update HISTORY.md and version.h for 8.3.3

Fix a bug where iterator can return incorrect data for DeleteRange() users (#11786 )
Summary: This should only affect iterator when - user uses DeleteRange(), - An iterator from level L has a non-ok status (such non-ok status may not be caught before the bug fix in https://github.com/facebook/rocksdb/pull/11783), and - A range tombstone covers a key from level > L and triggers a reseek sets the status_ to OK in SeekImpl()/SeekPrevImpl() e.g. https://github.com/facebook/rocksdb/blob/bd6a8340c3a2db764620e90b3ac5be173fc68a0c/table/merging_iterator.cc#L801 Pull Request resolved: https://github.com/facebook/rocksdb/pull/11786 Differential Revision: D48908830 Pulled By: cbi42 fbshipit-source-id: eb564be375af4e33dc27542eff753260186e6d5d
2026-11-06 11:02:36 -05:00 · 2023-09-01 13:45:29 -07:00 · 2023-09-01 12:08:33 -07:00 · 2023-09-01 12:08:20 -07:00 · 2023-09-01 12:08:11 -07:00 · 2023-06-14 22:32:14 -07:00
392 changed files with 7764 additions and 4152 deletions
@@ -95,3 +95,4 @@ fuzz/crash-*

 cmake-build-*
 third-party/folly/
+.cache
@@ -253,33 +253,10 @@ if(CMAKE_SYSTEM_PROCESSOR MATCHES "loongarch64")
  endif(HAS_LOONGARCH64)
 endif(CMAKE_SYSTEM_PROCESSOR MATCHES "loongarch64")

-option(PORTABLE "build a portable binary" OFF)
-option(FORCE_SSE42 "force building with SSE4.2, even when PORTABLE=ON" OFF)
-option(FORCE_AVX "force building with AVX, even when PORTABLE=ON" OFF)
-option(FORCE_AVX2 "force building with AVX2, even when PORTABLE=ON" OFF)
-if(PORTABLE)
-  add_definitions(-DROCKSDB_PORTABLE)
-
-  # MSVC does not need a separate compiler flag to enable SSE4.2; if nmmintrin.h
-  # is available, it is available by default.
-  if(FORCE_SSE42 AND NOT MSVC)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse4.2 -mpclmul")
-  endif()
-  if(MSVC)
-    if(FORCE_AVX)
-      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX")
-    endif()
-    # MSVC automatically enables BMI / lzcnt with AVX2.
-    if(FORCE_AVX2)
-      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
-    endif()
-  else()
-    if(FORCE_AVX)
-      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx")
-    endif()
-    if(FORCE_AVX2)
-      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mavx2 -mbmi -mlzcnt")
-    endif()
+set(PORTABLE 0 CACHE STRING "Minimum CPU arch to support, or 0 = current CPU, 1 = baseline CPU")
+if(PORTABLE STREQUAL 1)
+  # Usually nothing to do; compiler default is typically the most general
+  if(NOT MSVC)
    if(CMAKE_SYSTEM_PROCESSOR MATCHES "^s390x")
      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=z196")
    endif()
@@ -287,10 +264,21 @@ if(PORTABLE)
      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=loongarch64")
    endif()
  endif()
+elseif(PORTABLE MATCHES [^0]+)
+  # Name of a CPU arch spec or feature set to require
+  if(MSVC)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:${PORTABLE}")
+  else()
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=${PORTABLE}")
+  endif()
 else()
  if(MSVC)
+    # NOTE: No auto-detection of current CPU, but instead assume some useful
+    # level of optimization is supported
    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
  else()
+    # Require instruction set from current CPU (with some legacy or opt-out
+    # exceptions)
    if(CMAKE_SYSTEM_PROCESSOR MATCHES "^s390x" AND NOT HAS_S390X_MARCH_NATIVE)
      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=z196")
    elseif(NOT CMAKE_SYSTEM_PROCESSOR MATCHES "^(powerpc|ppc)64" AND NOT HAS_ARMV8_CRC)
@@ -305,25 +293,6 @@ if(NOT MSVC)
  set(CMAKE_REQUIRED_FLAGS "-msse4.2 -mpclmul")
 endif()

-CHECK_CXX_SOURCE_COMPILES("
-#include <cstdint>
-#include <nmmintrin.h>
-#include <wmmintrin.h>
-int main() {
-  volatile uint32_t x = _mm_crc32_u32(0, 0);
-  const auto a = _mm_set_epi64x(0, 0);
-  const auto b = _mm_set_epi64x(0, 0);
-  const auto c = _mm_clmulepi64_si128(a, b, 0x00);
-  auto d = _mm_cvtsi128_si64(c);
-}
-" HAVE_SSE42)
-if(HAVE_SSE42)
-  add_definitions(-DHAVE_SSE42)
-  add_definitions(-DHAVE_PCLMUL)
-elseif(FORCE_SSE42)
-  message(FATAL_ERROR "FORCE_SSE42=ON but unable to compile with SSE4.2 enabled")
-endif()
-
 # Check if -latomic is required or not
 if (NOT MSVC)
  set(CMAKE_REQUIRED_FLAGS "--std=c++17")
@@ -1010,12 +979,6 @@ if ( ROCKSDB_PLUGINS )
  endforeach()
 endif()

-if(HAVE_SSE42 AND NOT MSVC)
-  set_source_files_properties(
-    util/crc32c.cc
-    PROPERTIES COMPILE_FLAGS "-msse4.2 -mpclmul")
-endif()
-
 if(CMAKE_SYSTEM_PROCESSOR MATCHES "^(powerpc|ppc)64")
  list(APPEND SOURCES
    util/crc32c_ppc.c
@@ -1327,6 +1290,7 @@ if(WITH_TESTS)
        db/db_bloom_filter_test.cc
        db/db_compaction_filter_test.cc
        db/db_compaction_test.cc
+        db/db_clip_test.cc
        db/db_dynamic_level_test.cc
        db/db_encryption_test.cc
        db/db_flush_test.cc
@@ -1,6 +1,42 @@
 # Rocksdb Change Log
-## 8.2.1 (05/04/2023)
-* Public API comments and a Meta-internal fix to the buck build
+## 8.3.3 (09/01/2023)
+### Bug Fixes
+* Fix a bug where if there is an error reading from offset 0 of a file from L1+ and that the file is not the first file in the sorted run, data can be lost in compaction and read/scan can return incorrect results.
+* Fix a bug where iterator may return incorrect result for DeleteRange() users if there was an error reading from a file.
+* Fixed a race condition in `GenericRateLimiter` that could cause it to stop granting requests
+
+## 8.3.2 (06/14/2023)
+### Bug Fixes
+* Reduced cases of illegally using Env::Default() during static destruction by never destroying the internal PosixEnv itself (except for builds checking for memory leaks). (#11538)
+
+## 8.3.1 (06/07/2023)
+### Performance Improvements
+* Fixed higher read QPS during DB::Open() reading files created prior to #11406, especially when reading many small file (size < 52 MB) during DB::Open() and partitioned filter or index is used.
+
+## 8.3.0 (05/19/2023)
+### New Features
+* Introduced a new option `block_protection_bytes_per_key`, which can be used to enable per key-value integrity protection for in-memory blocks in block cache (#11287).
+* Added `JemallocAllocatorOptions::num_arenas`. Setting `num_arenas > 1` may mitigate mutex contention in the allocator, particularly in scenarios where block allocations commonly bypass jemalloc tcache.
+* Improve the operational safety of publishing a DB or SST files to many hosts by using different block cache hash seeds on different hosts. The exact behavior is controlled by new option `ShardedCacheOptions::hash_seed`, which also documents the solved problem in more detail.
+* Introduced a new option `CompactionOptionsFIFO::file_temperature_age_thresholds` that allows FIFO compaction to compact files to different temperatures based on key age (#11428).
+* Added a new ticker stat to count how many times RocksDB detected a corruption while verifying a block checksum: `BLOCK_CHECKSUM_MISMATCH_COUNT`.
+* New statistics `rocksdb.file.read.db.open.micros` that measures read time of block-based SST tables or blob files during db open.
+* New statistics tickers for various iterator seek behaviors and relevant filtering, as \*`_LEVEL_SEEK_`\*. (#11460)
+
+### Public API Changes
+* EXPERIMENTAL: Add new API `DB::ClipColumnFamily` to clip the key in CF to a certain range. It will physically deletes all keys outside the range including tombstones.
+* Add `MakeSharedCache()` construction functions to various cache Options objects, and deprecated the `NewWhateverCache()` functions with long parameter lists.
+* Changed the meaning of various Bloom filter stats (prefix vs. whole key), with iterator-related filtering only being tracked in the new \*`_LEVEL_SEEK_`\*. stats. (#11460)
+
+### Behavior changes
+* For x86, CPU features are no longer detected at runtime nor in build scripts, but in source code using common preprocessor defines. This will likely unlock some small performance improvements on some newer hardware, but could hurt performance of the kCRC32c checksum, which is no longer the default, on some "portable" builds. See PR #11419 for details.
+
+### Bug Fixes
+* Delete an empty WAL file on DB open if the log number is less than the min log number to keep
+* Delete temp OPTIONS file on DB open if there is a failure to write it out or rename it
+
+### Performance Improvements
+* Improved the I/O efficiency of prefetching SST metadata by recording more information in the DB manifest. Opening files written with previous versions will still rely on heuristics for how much to prefetch (#11406).

 ## 8.2.0 (04/24/2023)
 ### Public API Changes
@@ -20,12 +20,15 @@ There are few options when compiling RocksDB:
 depend on gflags. You will need to have gflags installed to run `make all`. This will compile RocksDB in debug mode. Don't
 use binaries compiled by `make all` in production.

-* By default the binary we produce is optimized for the platform you're compiling on
-(`-march=native` or the equivalent). SSE4.2 will thus be enabled automatically if your
-CPU supports it. To print a warning if your CPU does not support SSE4.2, build with
-`USE_SSE=1 make static_lib` or, if using CMake, `cmake -DFORCE_SSE42=ON`. If you want
-to build a portable binary, add `PORTABLE=1` before your make commands, like this:
-`PORTABLE=1 make static_lib`.
+* By default the binary we produce is optimized for the CPU you're compiling on
+(`-march=native` or the equivalent). To build a binary compatible with the most
+general architecture supported by your CPU and compiler, set `PORTABLE=1` for
+the build, but performance will suffer as many operations benefit from newer
+and wider instructions. In addition to `PORTABLE=0` (default) and `PORTABLE=1`,
+it can be set to an architecture name recognized by your compiler. For example,
+on 64-bit x86, a reasonable compromise is `PORTABLE=haswell` which supports
+many or most of the available optimizations while still being compatible with
+most processors made since roughly 2013.

 ## Dependencies

@@ -337,8 +337,8 @@ ifneq ($(MACHINE), arm64)
 # linking with jemalloc (as it won't be arm64-compatible) and remove some other options
 # set during platform detection
 DISABLE_JEMALLOC=1
-PLATFORM_CCFLAGS := $(filter-out -march=native -DHAVE_SSE42 -DHAVE_AVX2, $(PLATFORM_CCFLAGS))
-PLATFORM_CXXFLAGS := $(filter-out -march=native -DHAVE_SSE42 -DHAVE_AVX2, $(PLATFORM_CXXFLAGS))
+PLATFORM_CCFLAGS := $(filter-out -march=native, $(PLATFORM_CCFLAGS))
+PLATFORM_CXXFLAGS := $(filter-out -march=native, $(PLATFORM_CXXFLAGS))
 endif
 endif
 endif
@@ -1480,6 +1480,9 @@ db_compaction_filter_test: $(OBJ_DIR)/db/db_compaction_filter_test.o $(TEST_LIBR
 db_compaction_test: $(OBJ_DIR)/db/db_compaction_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)

+db_clip_test: $(OBJ_DIR)/db/db_clip_test.o $(TEST_LIBRARY) $(LIBRARY)
+	$(AM_LINK)
+
 db_dynamic_level_test: $(OBJ_DIR)/db/db_dynamic_level_test.o $(TEST_LIBRARY) $(LIBRARY)
 	$(AM_LINK)

@@ -2435,6 +2438,8 @@ checkout_folly:
 	@# NOTE: this hack is required for gcc in some cases
 	perl -pi -e 's/(__has_include.<experimental.memory_resource>.)/__cpp_rtti && $$1/' third-party/folly/folly/memory/MemoryResource.h

+CXX_M_FLAGS = $(filter -m%, $(CXXFLAGS))
+
 build_folly:
 	FOLLY_INST_PATH=`cd third-party/folly; $(PYTHON) build/fbcode_builder/getdeps.py show-inst-dir`; \
 	if [ "$$FOLLY_INST_PATH" ]; then \
@@ -2445,8 +2450,8 @@ build_folly:
 	fi
 	# Restore the original version of Invoke.h with boost dependency
 	cd third-party/folly && ${GIT_COMMAND} checkout folly/functional/Invoke.h
-	cd third-party/folly && MAYBE_AVX2=`echo $(CXXFLAGS) | grep -o -- -DHAVE_AVX2 | sed 's/-DHAVE_AVX2/-mavx2/g' || true` && \
-		CXXFLAGS=" $$MAYBE_AVX2 -DHAVE_CXX11_ATOMIC " $(PYTHON) build/fbcode_builder/getdeps.py build --no-tests
+	cd third-party/folly && \
+		CXXFLAGS=" $(CXX_M_FLAGS) -DHAVE_CXX11_ATOMIC " $(PYTHON) build/fbcode_builder/getdeps.py build --no-tests

 # ---------------------------------------------------------------------------
 #   Build size testing
@@ -5,3 +5,4 @@ This is the list of all known third-party plugins for RocksDB. If something is m
 * [ZenFS](https://github.com/westerndigitalcorporation/zenfs): a file system for zoned block devices
 * [RADOS](https://github.com/riversand963/rocksdb-rados-env): an Env used for interacting with RADOS. Migrated from RocksDB main repo.
 * [PMEM](https://github.com/pmem/pmem-rocksdb-plugin): a collection of plugins to enable Persistent Memory on RocksDB.
+* [IPPCP](https://github.com/intel/ippcp-plugin-rocksdb): a plugin to enable encryption on RocksDB based on Intel optimized open source IPP-Crypto library.
@@ -392,6 +392,7 @@ rocks_cpp_library_wrapper(name="rocksdb_stress_lib", srcs=[
        "db_stress_tool/db_stress_test_base.cc",
        "db_stress_tool/db_stress_tool.cc",
        "db_stress_tool/expected_state.cc",
+        "db_stress_tool/expected_value.cc",
        "db_stress_tool/multi_ops_txns_stress.cc",
        "db_stress_tool/no_batched_ops_stress.cc",
        "test_util/testutil.cc",
@@ -4749,6 +4750,12 @@ cpp_unittest_wrapper(name="db_bloom_filter_test",
            extra_compiler_flags=[])


+cpp_unittest_wrapper(name="db_clip_test",
+            srcs=["db/db_clip_test.cc"],
+            deps=[":rocksdb_test_lib"],
+            extra_compiler_flags=[])
+
+
 cpp_unittest_wrapper(name="db_compaction_filter_test",
            srcs=["db/db_compaction_filter_test.cc"],
            deps=[":rocksdb_test_lib"],
@@ -15,6 +15,25 @@ At Facebook, we use RocksDB as storage engines in multiple data management servi

 [2] https://code.facebook.com/posts/357056558062811/logdevice-a-distributed-data-store-for-logs/

+## Bilibili
+[Bilibili](bilibili.com) [uses](https://www.alluxio.io/blog/when-ai-meets-alluxio-at-bilibili-building-an-efficient-ai-platform-for-data-preprocessing-and-model-training/) Alluxio to speed up its ML training workloads, and Alluxio uses RocksDB to store its filesystem metadata, so Bilibili uses RocksDB.
+
+Bilibili's [real-time platform](https://www.alibabacloud.com/blog/architecture-and-practices-of-bilibilis-real-time-platform_596676) uses Flink, and uses RocksDB as Flink's state store.
+
+## TikTok
+TikTok, or its parent company ByteDance, uses RocksDB as the storage engine for some storage systems, such as its distributed graph database [ByteGraph](https://vldb.org/pvldb/vol15/p3306-li.pdf). 
+
+Also, TikTok uses [Alluxio](alluxio.io) to [speed up Presto queries](https://www.alluxio.io/resources/videos/improving-presto-performance-with-alluxio-at-tiktok/), and Alluxio stores the files' metadata in RocksDB.
+
+## FoundationDB
+[FoundationDB](https://www.foundationdb.org/) [uses](https://github.com/apple/foundationdb/blob/377f1f692da6ab2fe5bdac57035651db3e5fb66d/fdbserver/KeyValueStoreRocksDB.actor.cpp) RocksDB to implement a [key-value store interface](https://github.com/apple/foundationdb/blob/377f1f692da6ab2fe5bdac57035651db3e5fb66d/fdbserver/KeyValueStoreRocksDB.actor.cpp#L1127) in its server backend.
+
+## Apple
+Apple [uses](https://opensource.apple.com/projects/foundationdb/) FoundationDB, so it also uses RocksDB.
+
+## Snowflake
+Snowflake [uses](https://www.snowflake.com/blog/how-foundationdb-powers-snowflake-metadata-forward/) FoundationDB, so it also uses RocksDB.
+
 ## Microsoft
 The Bing search engine from Microsoft uses RocksDB as the storage engine for its web data platform: https://blogs.bing.com/Engineering-Blog/october-2021/RocksDB-in-Microsoft-Bing

@@ -85,9 +104,18 @@ quasardb uses a heavily tuned RocksDB as its persistence layer.
 ## TiKV
 [TiKV](https://github.com/pingcap/tikv) is a GEO-replicated, high-performance, distributed, transactional key-value database. TiKV is powered by Rust and Raft. TiKV uses RocksDB as its persistence layer.

+## TiDB
+[TiDB](https://github.com/pingcap/tidb) uses the TiKV distributed key-value database, so it uses RocksDB.
+
+## PingCAP
+[PingCAP](https://www.pingcap.com/) is the company behind TiDB, its cloud database service uses RocksDB.
+
 ## Apache Spark
 [Spark Structured Streaming](https://docs.databricks.com/structured-streaming/rocksdb-state-store.html) uses RocksDB as the local state store.

+## Databricks
+[Databricks](https://www.databricks.com/) [replaces AWS RDS with TiDB](https://www.pingcap.com/case-study/how-databricks-tackles-the-scalability-limit-with-a-mysql-alternative/) for scalability, so it uses RocksDB.
+
 ## Apache Flink
 [Apache Flink](https://flink.apache.org/news/2016/03/08/release-1.0.0.html) uses RocksDB to store state locally on a machine.

@@ -621,7 +621,7 @@ EOF
  fi
 fi

-if test "0$PORTABLE" -eq 0; then
+if [ "$PORTABLE" == "" ] || [ "$PORTABLE" == 0 ]; then
  if test -n "`echo $TARGET_ARCHITECTURE | grep ^ppc64`"; then
    # Tune for this POWER processor, treating '+' models as base models
    POWER=`LD_SHOW_AUXV=1 /bin/true | grep AT_PLATFORM | grep -E -o power[0-9]+`
@@ -644,27 +644,26 @@ if test "0$PORTABLE" -eq 0; then
    COMMON_FLAGS="$COMMON_FLAGS -march=${RISC_ISA}"
  elif [ "$TARGET_OS" == "IOS" ]; then
    COMMON_FLAGS="$COMMON_FLAGS"
-  elif [ "$TARGET_OS" == "AIX" ] || [ "$TARGET_OS" == "SunOS" ]; then
-    # TODO: Not sure why we don't use -march=native on these OSes
-    if test "$USE_SSE"; then
-      TRY_SSE_ETC="1"
-    fi
  else
    COMMON_FLAGS="$COMMON_FLAGS -march=native "
  fi
 else
-  # PORTABLE=1
-  if test "$USE_SSE"; then
-    TRY_SSE_ETC="1"
-  fi
-
-  if test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
-    COMMON_FLAGS="$COMMON_FLAGS -march=z196 "
-  fi
-
-  if test -n "`echo $TARGET_ARCHITECTURE | grep ^riscv64`"; then
-    RISC_ISA=$(cat /proc/cpuinfo | grep isa | head -1 | cut --delimiter=: -f 2 | cut -b 2-)
-    COMMON_FLAGS="$COMMON_FLAGS -march=${RISC_ISA}"
+  # PORTABLE specified
+  if [ "$PORTABLE" == 1 ]; then
+    if test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
+      COMMON_FLAGS="$COMMON_FLAGS -march=z196 "
+    elif test -n "`echo $TARGET_ARCHITECTURE | grep ^riscv64`"; then
+      RISC_ISA=$(cat /proc/cpuinfo | grep isa | head -1 | cut --delimiter=: -f 2 | cut -b 2-)
+      COMMON_FLAGS="$COMMON_FLAGS -march=${RISC_ISA}"
+    elif test "$USE_SSE"; then
+      # USE_SSE is DEPRECATED
+      # This is a rough approximation of the old USE_SSE behavior
+      COMMON_FLAGS="$COMMON_FLAGS -march=haswell"
+    fi
+    # Other than those cases, not setting -march= here.
+  else
+    # Assume PORTABLE is a minimum assumed cpu type, e.g. PORTABLE=haswell
+    COMMON_FLAGS="$COMMON_FLAGS -march=${PORTABLE}"
  fi

  if [[ "${PLATFORM}" == "OS_MACOSX" ]]; then
@@ -698,101 +697,6 @@ EOF
  fi
 fi

-if test "$TRY_SSE_ETC"; then
-  # The USE_SSE flag now means "attempt to compile with widely-available
-  # Intel architecture extensions utilized by specific optimizations in the
-  # source code." It's a qualifier on PORTABLE=1 that means "mostly portable."
-  # It doesn't even really check that your current CPU is compatible.
-  #
-  # SSE4.2 available since nehalem, ca. 2008-2010
-  # Includes POPCNT for BitsSetToOne, BitParity
-  TRY_SSE42="-msse4.2"
-  # PCLMUL available since westmere, ca. 2010-2011
-  TRY_PCLMUL="-mpclmul"
-  # AVX2 available since haswell, ca. 2013-2015
-  TRY_AVX2="-mavx2"
-  # BMI available since haswell, ca. 2013-2015
-  # Primarily for TZCNT for CountTrailingZeroBits
-  TRY_BMI="-mbmi"
-  # LZCNT available since haswell, ca. 2013-2015
-  # For FloorLog2
-  TRY_LZCNT="-mlzcnt"
-fi
-
-$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_SSE42 -x c++ - -o test.o 2>/dev/null <<EOF
-  #include <cstdint>
-  #include <nmmintrin.h>
-  int main() {
-    volatile uint32_t x = _mm_crc32_u32(0, 0);
-    (void)x;
-  }
-EOF
-if [ "$?" = 0 ]; then
-  COMMON_FLAGS="$COMMON_FLAGS $TRY_SSE42 -DHAVE_SSE42"
-elif test "$USE_SSE"; then
-  echo "warning: USE_SSE specified but compiler could not use SSE intrinsics, disabling" >&2
-fi
-
-$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_PCLMUL -x c++ - -o test.o 2>/dev/null <<EOF
-  #include <cstdint>
-  #include <wmmintrin.h>
-  int main() {
-    const auto a = _mm_set_epi64x(0, 0);
-    const auto b = _mm_set_epi64x(0, 0);
-    const auto c = _mm_clmulepi64_si128(a, b, 0x00);
-    auto d = _mm_cvtsi128_si64(c);
-    (void)d;
-  }
-EOF
-if [ "$?" = 0 ]; then
-  COMMON_FLAGS="$COMMON_FLAGS $TRY_PCLMUL -DHAVE_PCLMUL"
-elif test "$USE_SSE"; then
-  echo "warning: USE_SSE specified but compiler could not use PCLMUL intrinsics, disabling" >&2
-fi
-
-$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_AVX2 -x c++ - -o test.o 2>/dev/null <<EOF
-  #include <cstdint>
-  #include <immintrin.h>
-  int main() {
-    const auto a = _mm256_setr_epi32(0, 1, 2, 3, 4, 7, 6, 5);
-    const auto b = _mm256_permutevar8x32_epi32(a, a);
-    (void)b;
-  }
-EOF
-if [ "$?" = 0 ]; then
-  COMMON_FLAGS="$COMMON_FLAGS $TRY_AVX2 -DHAVE_AVX2"
-elif test "$USE_SSE"; then
-  echo "warning: USE_SSE specified but compiler could not use AVX2 intrinsics, disabling" >&2
-fi
-
-$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_BMI -x c++ - -o test.o 2>/dev/null <<EOF
-  #include <cstdint>
-  #include <immintrin.h>
-  int main(int argc, char *argv[]) {
-    (void)argv;
-    return (int)_tzcnt_u64((uint64_t)argc);
-  }
-EOF
-if [ "$?" = 0 ]; then
-  COMMON_FLAGS="$COMMON_FLAGS $TRY_BMI -DHAVE_BMI"
-elif test "$USE_SSE"; then
-  echo "warning: USE_SSE specified but compiler could not use BMI intrinsics, disabling" >&2
-fi
-
-$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS $TRY_LZCNT -x c++ - -o test.o 2>/dev/null <<EOF
-  #include <cstdint>
-  #include <immintrin.h>
-  int main(int argc, char *argv[]) {
-    (void)argv;
-    return (int)_lzcnt_u64((uint64_t)argc);
-  }
-EOF
-if [ "$?" = 0 ]; then
-  COMMON_FLAGS="$COMMON_FLAGS $TRY_LZCNT -DHAVE_LZCNT"
-elif test "$USE_SSE"; then
-  echo "warning: USE_SSE specified but compiler could not use LZCNT intrinsics, disabling" >&2
-fi
-
 $CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS -x c++ - -o test.o 2>/dev/null <<EOF
  #include <cstdint>
  int main() {
@@ -147,7 +147,7 @@ else
 fi

 CFLAGS+=" $DEPS_INCLUDE"
-CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DHAVE_SSE42"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT"
 CXXFLAGS+=" $CFLAGS"

 EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS"
@@ -154,7 +154,7 @@ CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE/linux "
 CFLAGS+=" -isystem $KERNEL_HEADERS_INCLUDE "

 CFLAGS+=" $DEPS_INCLUDE"
-CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DHAVE_SSE42 -DROCKSDB_IOURING_PRESENT"
+CFLAGS+=" -DROCKSDB_PLATFORM_POSIX -DROCKSDB_LIB_IO_POSIX -DROCKSDB_FALLOCATE_PRESENT -DROCKSDB_MALLOC_USABLE_SIZE -DROCKSDB_RANGESYNC_PRESENT -DROCKSDB_SCHED_GETCPU_PRESENT -DROCKSDB_IOURING_PRESENT"
 CXXFLAGS+=" $CFLAGS"

 EXEC_LDFLAGS=" $SNAPPY_LIBS $ZLIB_LIBS $BZIP_LIBS $LZ4_LIBS $ZSTD_LIBS $GFLAGS_LIBS $NUMA_LIB $TBB_LIBS $LIBURING_LIBS $BENCHMARK_LIBS"
@@ -21,6 +21,7 @@
 #include "test_util/secondary_cache_test_util.h"
 #include "test_util/testharness.h"
 #include "util/coding.h"
+#include "util/hash_containers.h"
 #include "util/string_util.h"

 // HyperClockCache only supports 16-byte keys, so some of the tests
@@ -955,6 +956,97 @@ TEST_P(CacheTest, GetChargeAndDeleter) {
  cache_->Release(h1);
 }

+namespace {
+bool AreTwoCacheKeysOrdered(Cache* cache) {
+  std::vector<std::string> keys;
+  const auto callback = [&](const Slice& key, Cache::ObjectPtr /*value*/,
+                            size_t /*charge*/,
+                            const Cache::CacheItemHelper* /*helper*/) {
+    keys.push_back(key.ToString());
+  };
+  cache->ApplyToAllEntries(callback, /*opts*/ {});
+  EXPECT_EQ(keys.size(), 2U);
+  EXPECT_NE(keys[0], keys[1]);
+  return keys[0] < keys[1];
+}
+}  // namespace
+
+TEST_P(CacheTest, CacheUniqueSeeds) {
+  // kQuasiRandomHashSeed should generate unique seeds (up to 2 billion before
+  // repeating)
+  UnorderedSet<uint32_t> seeds_seen;
+  // Roughly sqrt(number of possible values) for a decent chance at detecting
+  // a random collision if it's possible (shouldn't be)
+  uint16_t kSamples = 20000;
+  seeds_seen.reserve(kSamples);
+
+  // Hash seed should affect ordering of entries in the table, so we should
+  // have extremely high chance of seeing two entries ordered both ways.
+  bool seen_forward_order = false;
+  bool seen_reverse_order = false;
+
+  for (int i = 0; i < kSamples; ++i) {
+    auto cache = NewCache(2, [=](ShardedCacheOptions& opts) {
+      opts.hash_seed = LRUCacheOptions::kQuasiRandomHashSeed;
+      opts.num_shard_bits = 0;
+      opts.metadata_charge_policy = kDontChargeCacheMetadata;
+    });
+    auto val = cache->GetHashSeed();
+    ASSERT_TRUE(seeds_seen.insert(val).second);
+
+    ASSERT_OK(cache->Insert(EncodeKey(1), nullptr, &kHelper, /*charge*/ 1));
+    ASSERT_OK(cache->Insert(EncodeKey(2), nullptr, &kHelper, /*charge*/ 1));
+
+    if (AreTwoCacheKeysOrdered(cache.get())) {
+      seen_forward_order = true;
+    } else {
+      seen_reverse_order = true;
+    }
+  }
+
+  ASSERT_TRUE(seen_forward_order);
+  ASSERT_TRUE(seen_reverse_order);
+}
+
+TEST_P(CacheTest, CacheHostSeed) {
+  // kHostHashSeed should generate a consistent seed within this process
+  // (and other processes on the same host, but not unit testing that).
+  // And we should be able to use that chosen seed as an explicit option
+  // (for debugging).
+  // And we should verify consistent ordering of entries.
+  uint32_t expected_seed = 0;
+  bool expected_order = false;
+  // 10 iterations -> chance of a random seed falsely appearing consistent
+  // should be low, just 1 in 2^9.
+  for (int i = 0; i < 10; ++i) {
+    auto cache = NewCache(2, [=](ShardedCacheOptions& opts) {
+      if (i != 5) {
+        opts.hash_seed = LRUCacheOptions::kHostHashSeed;
+      } else {
+        // Can be used as explicit seed
+        opts.hash_seed = static_cast<int32_t>(expected_seed);
+        ASSERT_GE(opts.hash_seed, 0);
+      }
+      opts.num_shard_bits = 0;
+      opts.metadata_charge_policy = kDontChargeCacheMetadata;
+    });
+    ASSERT_OK(cache->Insert(EncodeKey(1), nullptr, &kHelper, /*charge*/ 1));
+    ASSERT_OK(cache->Insert(EncodeKey(2), nullptr, &kHelper, /*charge*/ 1));
+    uint32_t val = cache->GetHashSeed();
+    bool order = AreTwoCacheKeysOrdered(cache.get());
+    if (i != 0) {
+      ASSERT_EQ(val, expected_seed);
+      ASSERT_EQ(order, expected_order);
+    } else {
+      expected_seed = val;
+      expected_order = order;
+    }
+  }
+  // Printed for reference in case it's needed to reproduce other unit test
+  // failures on another host
+  fprintf(stderr, "kHostHashSeed -> %u\n", (unsigned)expected_seed);
+}
+
 INSTANTIATE_TEST_CASE_P(CacheTestInstance, CacheTest,
                        secondary_cache_test_util::GetTestingCacheTypes());
 INSTANTIATE_TEST_CASE_P(CacheTestInstance, LRUCacheTest,
@@ -16,7 +16,7 @@
 #include "cache/secondary_cache_adapter.h"
 #include "logging/logging.h"
 #include "monitoring/perf_context_imp.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "port/lang.h"
 #include "rocksdb/env.h"
 #include "util/hash.h"
@@ -130,7 +130,8 @@ HyperClockTable::HyperClockTable(
    size_t capacity, bool /*strict_capacity_limit*/,
    CacheMetadataChargePolicy metadata_charge_policy,
    MemoryAllocator* allocator,
-    const Cache::EvictionCallback* eviction_callback, const Opts& opts)
+    const Cache::EvictionCallback* eviction_callback, const uint32_t* hash_seed,
+    const Opts& opts)
    : length_bits_(CalcHashBits(capacity, opts.estimated_value_size,
                                metadata_charge_policy)),
      length_bits_mask_((size_t{1} << length_bits_) - 1),
@@ -138,7 +139,8 @@ HyperClockTable::HyperClockTable(
                                           kStrictLoadFactor)),
      array_(new HandleImpl[size_t{1} << length_bits_]),
      allocator_(allocator),
-      eviction_callback_(*eviction_callback) {
+      eviction_callback_(*eviction_callback),
+      hash_seed_(*hash_seed) {
  if (metadata_charge_policy ==
      CacheMetadataChargePolicy::kFullChargeCacheMetadata) {
    usage_ += size_t{GetTableSize()} * sizeof(HandleImpl);
@@ -1010,7 +1012,7 @@ inline void HyperClockTable::Evict(size_t requested_charge,
        if (eviction_callback_) {
          took_ownership =
              eviction_callback_(ClockCacheShard<HyperClockTable>::ReverseHash(
-                                     h.GetHash(), &unhashed),
+                                     h.GetHash(), &unhashed, hash_seed_),
                                 reinterpret_cast<Cache::Handle*>(&h));
        }
        if (!took_ownership) {
@@ -1039,11 +1041,11 @@ ClockCacheShard<Table>::ClockCacheShard(
    size_t capacity, bool strict_capacity_limit,
    CacheMetadataChargePolicy metadata_charge_policy,
    MemoryAllocator* allocator,
-    const Cache::EvictionCallback* eviction_callback,
+    const Cache::EvictionCallback* eviction_callback, const uint32_t* hash_seed,
    const typename Table::Opts& opts)
    : CacheShardBase(metadata_charge_policy),
      table_(capacity, strict_capacity_limit, metadata_charge_policy, allocator,
-             eviction_callback, opts),
+             eviction_callback, hash_seed, opts),
      capacity_(capacity),
      strict_capacity_limit_(strict_capacity_limit) {
  // Initial charge metadata should not exceed capacity
@@ -1082,10 +1084,11 @@ void ClockCacheShard<Table>::ApplyToSomeEntries(
    *state = index_end << (sizeof(size_t) * 8u - length_bits);
  }

+  auto hash_seed = table_.GetHashSeed();
  table_.ConstApplyToEntriesRange(
-      [callback](const HandleImpl& h) {
+      [callback, hash_seed](const HandleImpl& h) {
        UniqueId64x2 unhashed;
-        callback(ReverseHash(h.hashed_key, &unhashed), h.value,
+        callback(ReverseHash(h.hashed_key, &unhashed, hash_seed), h.value,
                 h.GetTotalCharge(), h.helper);
      },
      index_begin, index_end, false);
@@ -1282,25 +1285,20 @@ size_t ClockCacheShard<Table>::GetTableAddressCount() const {
 // Explicit instantiation
 template class ClockCacheShard<HyperClockTable>;

-HyperClockCache::HyperClockCache(
-    size_t capacity, size_t estimated_value_size, int num_shard_bits,
-    bool strict_capacity_limit,
-    CacheMetadataChargePolicy metadata_charge_policy,
-    std::shared_ptr<MemoryAllocator> memory_allocator)
-    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit,
-                   std::move(memory_allocator)) {
-  assert(estimated_value_size > 0 ||
-         metadata_charge_policy != kDontChargeCacheMetadata);
+HyperClockCache::HyperClockCache(const HyperClockCacheOptions& opts)
+    : ShardedCache(opts) {
+  assert(opts.estimated_entry_charge > 0 ||
+         opts.metadata_charge_policy != kDontChargeCacheMetadata);
  // TODO: should not need to go through two levels of pointer indirection to
  // get to table entries
  size_t per_shard = GetPerShardCapacity();
  MemoryAllocator* alloc = this->memory_allocator();
-  const Cache::EvictionCallback* eviction_callback = &eviction_callback_;
-  InitShards([=](Shard* cs) {
-    HyperClockTable::Opts opts;
-    opts.estimated_value_size = estimated_value_size;
-    new (cs) Shard(per_shard, strict_capacity_limit, metadata_charge_policy,
-                   alloc, eviction_callback, opts);
+  InitShards([&](Shard* cs) {
+    HyperClockTable::Opts table_opts;
+    table_opts.estimated_value_size = opts.estimated_entry_charge;
+    new (cs) Shard(per_shard, opts.strict_capacity_limit,
+                   opts.metadata_charge_policy, alloc, &eviction_callback_,
+                   &hash_seed_, table_opts);
  });
 }

@@ -1460,21 +1458,23 @@ std::shared_ptr<Cache> NewClockCache(
 }

 std::shared_ptr<Cache> HyperClockCacheOptions::MakeSharedCache() const {
-  auto my_num_shard_bits = num_shard_bits;
-  if (my_num_shard_bits >= 20) {
+  // For sanitized options
+  HyperClockCacheOptions opts = *this;
+  if (opts.num_shard_bits >= 20) {
    return nullptr;  // The cache cannot be sharded into too many fine pieces.
  }
-  if (my_num_shard_bits < 0) {
+  if (opts.num_shard_bits < 0) {
    // Use larger shard size to reduce risk of large entries clustering
    // or skewing individual shards.
    constexpr size_t min_shard_size = 32U * 1024U * 1024U;
-    my_num_shard_bits = GetDefaultCacheShardBits(capacity, min_shard_size);
+    opts.num_shard_bits =
+        GetDefaultCacheShardBits(opts.capacity, min_shard_size);
  }
-  std::shared_ptr<Cache> cache = std::make_shared<clock_cache::HyperClockCache>(
-      capacity, estimated_entry_charge, my_num_shard_bits,
-      strict_capacity_limit, metadata_charge_policy, memory_allocator);
-  if (secondary_cache) {
-    cache = std::make_shared<CacheWithSecondaryAdapter>(cache, secondary_cache);
+  std::shared_ptr<Cache> cache =
+      std::make_shared<clock_cache::HyperClockCache>(opts);
+  if (opts.secondary_cache) {
+    cache = std::make_shared<CacheWithSecondaryAdapter>(cache,
+                                                        opts.secondary_cache);
  }
  return cache;
 }
@@ -407,7 +407,7 @@ class HyperClockTable {
                  CacheMetadataChargePolicy metadata_charge_policy,
                  MemoryAllocator* allocator,
                  const Cache::EvictionCallback* eviction_callback,
-                  const Opts& opts);
+                  const uint32_t* hash_seed, const Opts& opts);
  ~HyperClockTable();

  Status Insert(const ClockHandleBasicData& proto, HandleImpl** handle,
@@ -448,6 +448,8 @@ class HyperClockTable {
    return standalone_usage_.load(std::memory_order_relaxed);
  }

+  uint32_t GetHashSeed() const { return hash_seed_; }
+
  // Acquire/release N references
  void TEST_RefN(HandleImpl& handle, size_t n);
  void TEST_ReleaseN(HandleImpl* handle, size_t n);
@@ -546,6 +548,9 @@ class HyperClockTable {
  // A reference to Cache::eviction_callback_
  const Cache::EvictionCallback& eviction_callback_;

+  // A reference to ShardedCacheBase::hash_seed_
+  const uint32_t& hash_seed_;
+
  // We partition the following members into different cache lines
  // to avoid false sharing among Lookup, Release, Erase and Insert
  // operations in ClockCacheShard.
@@ -573,7 +578,7 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {
                  CacheMetadataChargePolicy metadata_charge_policy,
                  MemoryAllocator* allocator,
                  const Cache::EvictionCallback* eviction_callback,
-                  const typename Table::Opts& opts);
+                  const uint32_t* hash_seed, const typename Table::Opts& opts);

  // For CacheShard concept
  using HandleImpl = typename Table::HandleImpl;
@@ -583,22 +588,23 @@ class ALIGN_AS(CACHE_LINE_SIZE) ClockCacheShard final : public CacheShardBase {
  static inline uint32_t HashPieceForSharding(HashCref hash) {
    return Upper32of64(hash[0]);
  }
-  static inline HashVal ComputeHash(const Slice& key) {
+  static inline HashVal ComputeHash(const Slice& key, uint32_t seed) {
    assert(key.size() == kCacheKeySize);
    HashVal in;
    HashVal out;
    // NOTE: endian dependence
    // TODO: use GetUnaligned?
    std::memcpy(&in, key.data(), kCacheKeySize);
-    BijectiveHash2x64(in[1], in[0], &out[1], &out[0]);
+    BijectiveHash2x64(in[1], in[0] ^ seed, &out[1], &out[0]);
    return out;
  }

  // For reconstructing key from hashed_key. Requires the caller to provide
  // backing storage for the Slice in `unhashed`
  static inline Slice ReverseHash(const UniqueId64x2& hashed,
-                                  UniqueId64x2* unhashed) {
+                                  UniqueId64x2* unhashed, uint32_t seed) {
    BijectiveUnhash2x64(hashed[1], hashed[0], &(*unhashed)[1], &(*unhashed)[0]);
+    (*unhashed)[0] ^= seed;
    // NOTE: endian dependence
    return Slice(reinterpret_cast<const char*>(unhashed), kCacheKeySize);
  }
@@ -682,10 +688,7 @@ class HyperClockCache
 public:
  using Shard = ClockCacheShard<HyperClockTable>;

-  HyperClockCache(size_t capacity, size_t estimated_value_size,
-                  int num_shard_bits, bool strict_capacity_limit,
-                  CacheMetadataChargePolicy metadata_charge_policy,
-                  std::shared_ptr<MemoryAllocator> memory_allocator);
+  explicit HyperClockCache(const HyperClockCacheOptions& opts);

  const char* Name() const override { return "HyperClockCache"; }

@@ -9,7 +9,7 @@
 #include <cstdint>
 #include <memory>

-#include "memory/memory_allocator.h"
+#include "memory/memory_allocator_impl.h"
 #include "monitoring/perf_context_imp.h"
 #include "util/compression.h"
 #include "util/string_util.h"
@@ -17,23 +17,8 @@
 namespace ROCKSDB_NAMESPACE {

 CompressedSecondaryCache::CompressedSecondaryCache(
-    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-    double high_pri_pool_ratio, double low_pri_pool_ratio,
-    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
-    CacheMetadataChargePolicy metadata_charge_policy,
-    CompressionType compression_type, uint32_t compress_format_version,
-    bool enable_custom_split_merge,
-    const CacheEntryRoleSet& do_not_compress_roles)
-    : cache_options_(capacity, num_shard_bits, strict_capacity_limit,
-                     high_pri_pool_ratio, low_pri_pool_ratio, memory_allocator,
-                     use_adaptive_mutex, metadata_charge_policy,
-                     compression_type, compress_format_version,
-                     enable_custom_split_merge, do_not_compress_roles) {
-  cache_ =
-      NewLRUCache(capacity, num_shard_bits, strict_capacity_limit,
-                  high_pri_pool_ratio, memory_allocator, use_adaptive_mutex,
-                  metadata_charge_policy, low_pri_pool_ratio);
-}
+    const CompressedSecondaryCacheOptions& opts)
+    : cache_(opts.LRUCacheOptions::MakeSharedCache()), cache_options_(opts) {}

 CompressedSecondaryCache::~CompressedSecondaryCache() { cache_.reset(); }

@@ -311,31 +296,9 @@ const Cache::CacheItemHelper* CompressedSecondaryCache::GetHelper(
  }
 }

-std::shared_ptr<SecondaryCache> NewCompressedSecondaryCache(
-    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-    double high_pri_pool_ratio, double low_pri_pool_ratio,
-    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
-    CacheMetadataChargePolicy metadata_charge_policy,
-    CompressionType compression_type, uint32_t compress_format_version,
-    bool enable_custom_split_merge,
-    const CacheEntryRoleSet& do_not_compress_roles) {
-  return std::make_shared<CompressedSecondaryCache>(
-      capacity, num_shard_bits, strict_capacity_limit, high_pri_pool_ratio,
-      low_pri_pool_ratio, memory_allocator, use_adaptive_mutex,
-      metadata_charge_policy, compression_type, compress_format_version,
-      enable_custom_split_merge, do_not_compress_roles);
-}
-
-std::shared_ptr<SecondaryCache> NewCompressedSecondaryCache(
-    const CompressedSecondaryCacheOptions& opts) {
-  // The secondary_cache is disabled for this LRUCache instance.
-  assert(opts.secondary_cache == nullptr);
-  return NewCompressedSecondaryCache(
-      opts.capacity, opts.num_shard_bits, opts.strict_capacity_limit,
-      opts.high_pri_pool_ratio, opts.low_pri_pool_ratio, opts.memory_allocator,
-      opts.use_adaptive_mutex, opts.metadata_charge_policy,
-      opts.compression_type, opts.compress_format_version,
-      opts.enable_custom_split_merge, opts.do_not_compress_roles);
+std::shared_ptr<SecondaryCache>
+CompressedSecondaryCacheOptions::MakeSharedSecondaryCache() const {
+  return std::make_shared<CompressedSecondaryCache>(*this);
 }

 }  // namespace ROCKSDB_NAMESPACE
@@ -10,7 +10,7 @@
 #include <memory>

 #include "cache/lru_cache.h"
-#include "memory/memory_allocator.h"
+#include "memory/memory_allocator_impl.h"
 #include "rocksdb/secondary_cache.h"
 #include "rocksdb/slice.h"
 #include "rocksdb/status.h"
@@ -69,18 +69,8 @@ class CompressedSecondaryCacheResultHandle : public SecondaryCacheResultHandle {

 class CompressedSecondaryCache : public SecondaryCache {
 public:
-  CompressedSecondaryCache(
-      size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-      double high_pri_pool_ratio, double low_pri_pool_ratio,
-      std::shared_ptr<MemoryAllocator> memory_allocator = nullptr,
-      bool use_adaptive_mutex = kDefaultToAdaptiveMutex,
-      CacheMetadataChargePolicy metadata_charge_policy =
-          kDefaultCacheMetadataChargePolicy,
-      CompressionType compression_type = CompressionType::kLZ4Compression,
-      uint32_t compress_format_version = 2,
-      bool enable_custom_split_merge = false,
-      const CacheEntryRoleSet& do_not_compress_roles = {
-          CacheEntryRole::kFilterBlock});
+  explicit CompressedSecondaryCache(
+      const CompressedSecondaryCacheOptions& opts);
  ~CompressedSecondaryCache() override;

  const char* Name() const override { return "CompressedSecondaryCache"; }
@@ -626,8 +626,9 @@ class CompressedSecondaryCacheTestBase : public testing::Test,

    using CacheValueChunk = CompressedSecondaryCache::CacheValueChunk;
    std::unique_ptr<CompressedSecondaryCache> sec_cache =
-        std::make_unique<CompressedSecondaryCache>(1000, 0, true, 0.5, 0.0,
-                                                   allocator);
+        std::make_unique<CompressedSecondaryCache>(
+            CompressedSecondaryCacheOptions(1000, 0, true, 0.5, 0.0,
+                                            allocator));
    Random rnd(301);
    // 8500 = 8169 + 233 + 98, so there should be 3 chunks after split.
    size_t str_size{8500};
@@ -678,7 +679,8 @@ class CompressedSecondaryCacheTestBase : public testing::Test,
    std::string str = str1 + str2 + str3;

    std::unique_ptr<CompressedSecondaryCache> sec_cache =
-        std::make_unique<CompressedSecondaryCache>(1000, 0, true, 0.5, 0.0);
+        std::make_unique<CompressedSecondaryCache>(
+            CompressedSecondaryCacheOptions(1000, 0, true, 0.5, 0.0));
    size_t charge{0};
    CacheAllocationPtr value =
        sec_cache->MergeChunksIntoValue(chunks_head, charge);
@@ -708,8 +710,9 @@ class CompressedSecondaryCacheTestBase : public testing::Test,

    using CacheValueChunk = CompressedSecondaryCache::CacheValueChunk;
    std::unique_ptr<CompressedSecondaryCache> sec_cache =
-        std::make_unique<CompressedSecondaryCache>(1000, 0, true, 0.5, 0.0,
-                                                   allocator);
+        std::make_unique<CompressedSecondaryCache>(
+            CompressedSecondaryCacheOptions(1000, 0, true, 0.5, 0.0,
+                                            allocator));
    Random rnd(301);
    // 8500 = 8169 + 233 + 98, so there should be 3 chunks after split.
    size_t str_size{8500};
@@ -16,7 +16,7 @@

 #include "cache/secondary_cache_adapter.h"
 #include "monitoring/perf_context_imp.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "port/lang.h"
 #include "util/distributed_mutex.h"

@@ -646,23 +646,15 @@ void LRUCacheShard::AppendPrintableOptions(std::string& str) const {
  str.append(buffer);
 }

-LRUCache::LRUCache(size_t capacity, int num_shard_bits,
-                   bool strict_capacity_limit, double high_pri_pool_ratio,
-                   double low_pri_pool_ratio,
-                   std::shared_ptr<MemoryAllocator> allocator,
-                   bool use_adaptive_mutex,
-                   CacheMetadataChargePolicy metadata_charge_policy)
-    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit,
-                   std::move(allocator)) {
+LRUCache::LRUCache(const LRUCacheOptions& opts) : ShardedCache(opts) {
  size_t per_shard = GetPerShardCapacity();
  MemoryAllocator* alloc = memory_allocator();
-  const EvictionCallback* eviction_callback = &eviction_callback_;
-  InitShards([=](LRUCacheShard* cs) {
-    new (cs) LRUCacheShard(per_shard, strict_capacity_limit,
-                           high_pri_pool_ratio, low_pri_pool_ratio,
-                           use_adaptive_mutex, metadata_charge_policy,
-                           /* max_upper_hash_bits */ 32 - num_shard_bits, alloc,
-                           eviction_callback);
+  InitShards([&](LRUCacheShard* cs) {
+    new (cs) LRUCacheShard(per_shard, opts.strict_capacity_limit,
+                           opts.high_pri_pool_ratio, opts.low_pri_pool_ratio,
+                           opts.use_adaptive_mutex, opts.metadata_charge_policy,
+                           /* max_upper_hash_bits */ 32 - opts.num_shard_bits,
+                           alloc, &eviction_callback_);
  });
 }

@@ -692,13 +684,7 @@ double LRUCache::GetHighPriPoolRatio() {

 }  // namespace lru_cache

-std::shared_ptr<Cache> NewLRUCache(
-    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-    double high_pri_pool_ratio,
-    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
-    CacheMetadataChargePolicy metadata_charge_policy,
-    const std::shared_ptr<SecondaryCache>& secondary_cache,
-    double low_pri_pool_ratio) {
+std::shared_ptr<Cache> LRUCacheOptions::MakeSharedCache() const {
  if (num_shard_bits >= 20) {
    return nullptr;  // The cache cannot be sharded into too many fine pieces.
  }
@@ -714,36 +700,24 @@ std::shared_ptr<Cache> NewLRUCache(
    // Invalid high_pri_pool_ratio and low_pri_pool_ratio combination
    return nullptr;
  }
-  if (num_shard_bits < 0) {
-    num_shard_bits = GetDefaultCacheShardBits(capacity);
+  // For sanitized options
+  LRUCacheOptions opts = *this;
+  if (opts.num_shard_bits < 0) {
+    opts.num_shard_bits = GetDefaultCacheShardBits(capacity);
  }
-  std::shared_ptr<Cache> cache = std::make_shared<LRUCache>(
-      capacity, num_shard_bits, strict_capacity_limit, high_pri_pool_ratio,
-      low_pri_pool_ratio, std::move(memory_allocator), use_adaptive_mutex,
-      metadata_charge_policy);
+  std::shared_ptr<Cache> cache = std::make_shared<LRUCache>(opts);
  if (secondary_cache) {
    cache = std::make_shared<CacheWithSecondaryAdapter>(cache, secondary_cache);
  }
  return cache;
 }

-std::shared_ptr<Cache> NewLRUCache(const LRUCacheOptions& cache_opts) {
-  return NewLRUCache(cache_opts.capacity, cache_opts.num_shard_bits,
-                     cache_opts.strict_capacity_limit,
-                     cache_opts.high_pri_pool_ratio,
-                     cache_opts.memory_allocator, cache_opts.use_adaptive_mutex,
-                     cache_opts.metadata_charge_policy,
-                     cache_opts.secondary_cache, cache_opts.low_pri_pool_ratio);
-}
-
-std::shared_ptr<Cache> NewLRUCache(
-    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-    double high_pri_pool_ratio,
-    std::shared_ptr<MemoryAllocator> memory_allocator, bool use_adaptive_mutex,
-    CacheMetadataChargePolicy metadata_charge_policy,
-    double low_pri_pool_ratio) {
-  return NewLRUCache(capacity, num_shard_bits, strict_capacity_limit,
-                     high_pri_pool_ratio, memory_allocator, use_adaptive_mutex,
-                     metadata_charge_policy, nullptr, low_pri_pool_ratio);
+std::shared_ptr<GeneralCache> LRUCacheOptions::MakeSharedGeneralCache() const {
+  if (secondary_cache) {
+    // Not allowed for a GeneralCache
+    return nullptr;
+  }
+  // Works while GeneralCache is an alias for Cache
+  return MakeSharedCache();
 }
 }  // namespace ROCKSDB_NAMESPACE
@@ -279,8 +279,8 @@ class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard final : public CacheShardBase {
  using HashCref = uint32_t;

 public:  // Function definitions expected as parameter to ShardedCache
-  static inline HashVal ComputeHash(const Slice& key) {
-    return Lower32of64(GetSliceNPHash64(key));
+  static inline HashVal ComputeHash(const Slice& key, uint32_t seed) {
+    return Lower32of64(GetSliceNPHash64(key, seed));
  }

  // Separate from constructor so caller can easily make an array of LRUCache
@@ -446,12 +446,7 @@ class LRUCache
 #endif
    : public ShardedCache<LRUCacheShard> {
 public:
-  LRUCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-           double high_pri_pool_ratio, double low_pri_pool_ratio,
-           std::shared_ptr<MemoryAllocator> memory_allocator = nullptr,
-           bool use_adaptive_mutex = kDefaultToAdaptiveMutex,
-           CacheMetadataChargePolicy metadata_charge_policy =
-               kDontChargeCacheMetadata);
+  explicit LRUCache(const LRUCacheOptions& opts);
  const char* Name() const override { return "LRUCache"; }
  ObjectPtr Value(Handle* handle) override;
  size_t GetCharge(Handle* handle) const override;
@@ -397,7 +397,7 @@ class ClockCacheTest : public testing::Test {
    opts.estimated_value_size = 1;
    new (shard_)
        Shard(capacity, strict_capacity_limit, kDontChargeCacheMetadata,
-              /*allocator*/ nullptr, &eviction_callback_, opts);
+              /*allocator*/ nullptr, &eviction_callback_, &hash_seed_, opts);
  }

  Status Insert(const UniqueId64x2& hashed_key,
@@ -455,6 +455,7 @@ class ClockCacheTest : public testing::Test {

 private:
  Cache::EvictionCallback eviction_callback_;
+  uint32_t hash_seed_ = 0;
 };

 TEST_F(ClockCacheTest, Misc) {
@@ -13,20 +13,57 @@
 #include <cstdint>
 #include <memory>

+#include "env/unique_id_gen.h"
+#include "rocksdb/env.h"
 #include "util/hash.h"
 #include "util/math.h"
 #include "util/mutexlock.h"

 namespace ROCKSDB_NAMESPACE {
+namespace {
+// The generated seeds must fit in 31 bits so that
+// ShardedCacheOptions::hash_seed can be set to it explicitly, for
+// diagnostic/debugging purposes.
+constexpr uint32_t kSeedMask = 0x7fffffff;
+uint32_t DetermineSeed(int32_t hash_seed_option) {
+  if (hash_seed_option >= 0) {
+    // User-specified exact seed
+    return static_cast<uint32_t>(hash_seed_option);
+  }
+  static SemiStructuredUniqueIdGen gen;
+  if (hash_seed_option == ShardedCacheOptions::kHostHashSeed) {
+    std::string hostname;
+    Status s = Env::Default()->GetHostNameString(&hostname);
+    if (s.ok()) {
+      return GetSliceHash(hostname) & kSeedMask;
+    } else {
+      // Fall back on something stable within the process.
+      return static_cast<uint32_t>(gen.GetBaseUpper()) & kSeedMask;
+    }
+  } else {
+    // for kQuasiRandomHashSeed and fallback
+    uint32_t val = gen.GenerateNext<uint32_t>() & kSeedMask;
+    // Perform some 31-bit bijective transformations so that we get
+    // quasirandom, not just incrementing. (An incrementing seed from a
+    // random starting point would be fine, but hard to describe in a name.)
+    // See https://en.wikipedia.org/wiki/Quasirandom and using a murmur-like
+    // transformation here for our bijection in the lower 31 bits.
+    // See https://en.wikipedia.org/wiki/MurmurHash
+    val *= /*31-bit prime*/ 1150630961;
+    val ^= (val & kSeedMask) >> 17;
+    val *= /*31-bit prime*/ 1320603883;
+    return val & kSeedMask;
+  }
+}
+}  // namespace

-ShardedCacheBase::ShardedCacheBase(size_t capacity, int num_shard_bits,
-                                   bool strict_capacity_limit,
-                                   std::shared_ptr<MemoryAllocator> allocator)
-    : Cache(std::move(allocator)),
+ShardedCacheBase::ShardedCacheBase(const ShardedCacheOptions& opts)
+    : Cache(opts.memory_allocator),
      last_id_(1),
-      shard_mask_((uint32_t{1} << num_shard_bits) - 1),
-      strict_capacity_limit_(strict_capacity_limit),
-      capacity_(capacity) {}
+      shard_mask_((uint32_t{1} << opts.num_shard_bits) - 1),
+      hash_seed_(DetermineSeed(opts.hash_seed)),
+      strict_capacity_limit_(opts.strict_capacity_limit),
+      capacity_(opts.capacity) {}

 size_t ShardedCacheBase::ComputePerShardCapacity(size_t capacity) const {
  uint32_t num_shards = GetNumShards();
@@ -34,8 +34,8 @@ class CacheShardBase {
  std::string GetPrintableOptions() const { return ""; }
  using HashVal = uint64_t;
  using HashCref = uint64_t;
-  static inline HashVal ComputeHash(const Slice& key) {
-    return GetSliceNPHash64(key);
+  static inline HashVal ComputeHash(const Slice& key, uint32_t seed) {
+    return GetSliceNPHash64(key, seed);
  }
  static inline uint32_t HashPieceForSharding(HashCref hash) {
    return Lower32of64(hash);
@@ -89,9 +89,7 @@ class CacheShardBase {
 // Portions of ShardedCache that do not depend on the template parameter
 class ShardedCacheBase : public Cache {
 public:
-  ShardedCacheBase(size_t capacity, int num_shard_bits,
-                   bool strict_capacity_limit,
-                   std::shared_ptr<MemoryAllocator> memory_allocator);
+  explicit ShardedCacheBase(const ShardedCacheOptions& opts);
  virtual ~ShardedCacheBase() = default;

  int GetNumShardBits() const;
@@ -106,6 +104,8 @@ class ShardedCacheBase : public Cache {
  size_t GetUsage(Handle* handle) const override;
  std::string GetPrintableOptions() const override;

+  uint32_t GetHashSeed() const override { return hash_seed_; }
+
 protected:  // fns
  virtual void AppendPrintableOptions(std::string& str) const = 0;
  size_t GetPerShardCapacity() const;
@@ -114,6 +114,7 @@ class ShardedCacheBase : public Cache {
 protected:                        // data
  std::atomic<uint64_t> last_id_;  // For NewId
  const uint32_t shard_mask_;
+  const uint32_t hash_seed_;

  // Dynamic configuration parameters, guarded by config_mutex_
  bool strict_capacity_limit_;
@@ -134,10 +135,8 @@ class ShardedCache : public ShardedCacheBase {
  using HashCref = typename CacheShard::HashCref;
  using HandleImpl = typename CacheShard::HandleImpl;

-  ShardedCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-               std::shared_ptr<MemoryAllocator> allocator)
-      : ShardedCacheBase(capacity, num_shard_bits, strict_capacity_limit,
-                         allocator),
+  explicit ShardedCache(const ShardedCacheOptions& opts)
+      : ShardedCacheBase(opts),
        shards_(reinterpret_cast<CacheShard*>(port::cacheline_aligned_alloc(
            sizeof(CacheShard) * GetNumShards()))),
        destroy_shards_in_dtor_(false) {}
@@ -175,7 +174,7 @@ class ShardedCache : public ShardedCacheBase {
                size_t charge, Handle** handle = nullptr,
                Priority priority = Priority::LOW) override {
    assert(helper);
-    HashVal hash = CacheShard::ComputeHash(key);
+    HashVal hash = CacheShard::ComputeHash(key, hash_seed_);
    auto h_out = reinterpret_cast<HandleImpl**>(handle);
    return GetShard(hash).Insert(key, hash, obj, helper, charge, h_out,
                                 priority);
@@ -185,7 +184,7 @@ class ShardedCache : public ShardedCacheBase {
                           const CacheItemHelper* helper, size_t charge,
                           bool allow_uncharged) override {
    assert(helper);
-    HashVal hash = CacheShard::ComputeHash(key);
+    HashVal hash = CacheShard::ComputeHash(key, hash_seed_);
    HandleImpl* result = GetShard(hash).CreateStandalone(
        key, hash, obj, helper, charge, allow_uncharged);
    return reinterpret_cast<Handle*>(result);
@@ -195,14 +194,14 @@ class ShardedCache : public ShardedCacheBase {
                 CreateContext* create_context = nullptr,
                 Priority priority = Priority::LOW,
                 Statistics* stats = nullptr) override {
-    HashVal hash = CacheShard::ComputeHash(key);
+    HashVal hash = CacheShard::ComputeHash(key, hash_seed_);
    HandleImpl* result = GetShard(hash).Lookup(key, hash, helper,
                                               create_context, priority, stats);
    return reinterpret_cast<Handle*>(result);
  }

  void Erase(const Slice& key) override {
-    HashVal hash = CacheShard::ComputeHash(key);
+    HashVal hash = CacheShard::ComputeHash(key, hash_seed_);
    GetShard(hash).Erase(key, hash);
  }

@@ -7,7 +7,7 @@

 #include <memory>

-#include "memory/memory_allocator.h"
+#include "memory/memory_allocator_impl.h"
 #include "rocksdb/advanced_cache.h"
 #include "rocksdb/rocksdb_namespace.h"
 #include "rocksdb/slice.h"
@@ -12,7 +12,7 @@
 #include "db/blob/blob_log_format.h"
 #include "file/file_prefetch_buffer.h"
 #include "file/filename.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "options/cf_options.h"
 #include "rocksdb/file_system.h"
 #include "rocksdb/slice.h"
@@ -7,7 +7,7 @@
 #include "db/blob/blob_log_sequential_reader.h"

 #include "file/random_access_file_reader.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "util/stop_watch.h"

 namespace ROCKSDB_NAMESPACE {
@@ -10,7 +10,7 @@

 #include "db/blob/blob_log_format.h"
 #include "file/writable_file_writer.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "rocksdb/system_clock.h"
 #include "test_util/sync_point.h"
 #include "util/coding.h"
@@ -13,7 +13,7 @@
 #include "db/blob/blob_contents.h"
 #include "db/blob/blob_file_reader.h"
 #include "db/blob/blob_log_format.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "options/cf_options.h"
 #include "table/get_context.h"
 #include "table/multiget_context.h"
@@ -291,6 +291,7 @@ Status BuildTable(
    if (s.ok() && !empty) {
      uint64_t file_size = builder->FileSize();
      meta->fd.file_size = file_size;
+      meta->tail_size = builder->GetTailSize();
      meta->marked_for_compaction = builder->NeedCompact();
      assert(meta->fd.GetFileSize() > 0);
      tp = builder
@@ -380,7 +381,8 @@ Status BuildTable(
          MaxFileSizeForL0MetaPin(mutable_cf_options),
          /*smallest_compaction_key=*/nullptr,
          /*largest_compaction_key*/ nullptr,
-          /*allow_unprepared_value*/ false));
+          /*allow_unprepared_value*/ false,
+          mutable_cf_options.block_protection_bytes_per_key));
      s = it->status();
      if (s.ok() && paranoid_file_checks) {
        OutputValidator file_validator(tboptions.internal_comparator,
@@ -1428,6 +1428,37 @@ Status ColumnFamilyData::ValidateOptions(
        "Memtable per key-value checksum protection only supports 0, 1, 2, 4 "
        "or 8 bytes per key.");
  }
+  if (std::find(supported.begin(), supported.end(),
+                cf_options.block_protection_bytes_per_key) == supported.end()) {
+    return Status::NotSupported(
+        "Block per key-value checksum protection only supports 0, 1, 2, 4 "
+        "or 8 bytes per key.");
+  }
+
+  if (!cf_options.compaction_options_fifo.file_temperature_age_thresholds
+           .empty()) {
+    if (cf_options.compaction_style != kCompactionStyleFIFO) {
+      return Status::NotSupported(
+          "Option file_temperature_age_thresholds only supports FIFO "
+          "compaction.");
+    } else if (cf_options.num_levels > 1) {
+      return Status::NotSupported(
+          "Option file_temperature_age_thresholds is only supported when "
+          "num_levels = 1.");
+    } else {
+      const auto& ages =
+          cf_options.compaction_options_fifo.file_temperature_age_thresholds;
+      assert(ages.size() >= 1);
+      // check that age is sorted
+      for (size_t i = 0; i < ages.size() - 1; ++i) {
+        if (ages[i].age >= ages[i + 1].age) {
+          return Status::NotSupported(
+              "Option file_temperature_age_thresholds requires elements to be "
+              "sorted in increasing order with respect to `age` field.");
+        }
+      }
+    }
+  }
  return s;
 }

@@ -465,6 +465,11 @@ bool Compaction::IsTrivialMove() const {
    return false;
  }

+  if (compaction_reason_ == CompactionReason::kChangeTemperature) {
+    // Changing temperature usually requires rewriting the file.
+    return false;
+  }
+
  // Used in universal compaction, where trivial move can be done if the
  // input files are non overlapping
  if ((mutable_cf_options_.compaction_options_universal.allow_trivial_move) &&
@@ -504,7 +504,9 @@ void CompactionJob::GenSubcompactionBoundaries() {
        FileMetaData* f = flevel->files[i].file_metadata;
        std::vector<TableReader::Anchor> my_anchors;
        Status s = cfd->table_cache()->ApproximateKeyAnchors(
-            read_options, icomp, *f, my_anchors);
+            read_options, icomp, *f,
+            c->mutable_cf_options()->block_protection_bytes_per_key,
+            my_anchors);
        if (!s.ok() || my_anchors.empty()) {
          my_anchors.emplace_back(f->largest.user_key(), f->fd.GetFileSize());
        }
@@ -735,7 +737,9 @@ Status CompactionJob::Run() {
                *compact_->compaction->mutable_cf_options()),
            /*smallest_compaction_key=*/nullptr,
            /*largest_compaction_key=*/nullptr,
-            /*allow_unprepared_value=*/false);
+            /*allow_unprepared_value=*/false,
+            compact_->compaction->mutable_cf_options()
+                ->block_protection_bytes_per_key);
        auto s = iter->status();

        if (s.ok() && paranoid_file_checks_) {
@@ -24,7 +24,7 @@
 #include "db/write_batch_internal.h"
 #include "env/mock_env.h"
 #include "file/filename.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "monitoring/thread_status_util.h"
 #include "port/stack_trace.h"
 #include "rocksdb/cache.h"
@@ -54,7 +54,7 @@
 #include "util/compression.h"
 #include "util/hash.h"
 #include "util/mutexlock.h"
-#include "util/rate_limiter.h"
+#include "util/rate_limiter_impl.h"
 #include "util/string_util.h"
 #include "utilities/merge_operators.h"

@@ -373,7 +373,7 @@ class CompactionJobTestBase : public testing::Test {
    } else if (table_type_ == TableTypeForTest::kMockTable) {
      file_size = 10;
      EXPECT_OK(mock_table_factory_->CreateMockTable(
-          env_, GenerateFileName(file_number), std::move(contents)));
+          env_, GenerateFileName(file_number), contents));
    } else {
      assert(false);
    }
@@ -386,7 +386,7 @@ class CompactionJobTestBase : public testing::Test {
        kUnknownFileCreationTime,
        versions_->GetColumnFamilySet()->GetDefault()->NewEpochNumber(),
        kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2,
-        0);
+        0, 0);

    mutex_.Lock();
    EXPECT_OK(versions_->LogAndApply(
@@ -454,7 +454,8 @@ class CompactionJobTestBase : public testing::Test {
      Status s = cf_options_.table_factory->NewTableReader(
          read_opts,
          TableReaderOptions(*cfd->ioptions(), nullptr, FileOptions(),
-                             cfd_->internal_comparator()),
+                             cfd_->internal_comparator(),
+                             0 /* block_protection_bytes_per_key */),
          std::move(freader), file_size, &table_reader, false);
      ASSERT_OK(s);
      assert(table_reader);
@@ -43,6 +43,7 @@ Status CompactionOutputs::Finish(const Status& intput_status,
  const uint64_t current_bytes = builder_->FileSize();
  if (s.ok()) {
    meta->fd.file_size = current_bytes;
+    meta->tail_size = builder_->GetTailSize();
    meta->marked_for_compaction = builder_->NeedCompact();
  }
  current_output().finished = true;
@@ -20,7 +20,7 @@
 #include "file/filename.h"
 #include "logging/log_buffer.h"
 #include "logging/logging.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "test_util/sync_point.h"
 #include "util/random.h"
 #include "util/string_util.h"
@@ -16,6 +16,7 @@
 #include "db/column_family.h"
 #include "logging/log_buffer.h"
 #include "logging/logging.h"
+#include "options/options_helper.h"
 #include "util/string_util.h"

 namespace ROCKSDB_NAMESPACE {
@@ -284,31 +285,36 @@ Compaction* FIFOCompactionPicker::PickSizeCompaction(
  return c;
 }

-Compaction* FIFOCompactionPicker::PickCompactionToWarm(
+Compaction* FIFOCompactionPicker::PickTemperatureChangeCompaction(
    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
    const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
    LogBuffer* log_buffer) {
-  if (mutable_cf_options.compaction_options_fifo.age_for_warm == 0) {
+  const std::vector<FileTemperatureAge>& ages =
+      mutable_cf_options.compaction_options_fifo
+          .file_temperature_age_thresholds;
+  if (ages.empty()) {
    return nullptr;
  }

-  // PickCompactionToWarm is only triggered if there is no non-L0 files.
-  for (int level = 1; level < vstorage->num_levels(); ++level) {
-    if (GetTotalFilesSize(vstorage->LevelFiles(level)) > 0) {
-      return nullptr;
-    }
+  // Does not apply to multi-level FIFO.
+  if (vstorage->num_levels() > 1) {
+    return nullptr;
  }

  const int kLevel0 = 0;
  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
+  if (level_files.empty()) {
+    return nullptr;
+  }

  int64_t _current_time;
  auto status = ioptions_.clock->GetCurrentTime(&_current_time);
  if (!status.ok()) {
-    ROCKS_LOG_BUFFER(log_buffer,
-                     "[%s] FIFO compaction: Couldn't get current time: %s. "
-                     "Not doing compactions based on warm threshold. ",
-                     cf_name.c_str(), status.ToString().c_str());
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] FIFO compaction: Couldn't get current time: %s. "
+        "Not doing compactions based on file temperature-age threshold. ",
+        cf_name.c_str(), status.ToString().c_str());
    return nullptr;
  }
  const uint64_t current_time = static_cast<uint64_t>(_current_time);
@@ -327,56 +333,77 @@ Compaction* FIFOCompactionPicker::PickCompactionToWarm(
  inputs[0].level = 0;

  // avoid underflow
-  if (current_time > mutable_cf_options.compaction_options_fifo.age_for_warm) {
-    uint64_t create_time_threshold =
-        current_time - mutable_cf_options.compaction_options_fifo.age_for_warm;
+  uint64_t min_age = ages[0].age;
+  // kLastTemperature means target temperature is to be determined.
+  Temperature compaction_target_temp = Temperature::kLastTemperature;
+  if (current_time > min_age) {
+    uint64_t create_time_threshold = current_time - min_age;
    uint64_t compaction_size = 0;
-    // We will ideally identify a file qualifying for warm tier by knowing
-    // the timestamp for the youngest entry in the file. However, right now
-    // we don't have the information. We infer it by looking at timestamp
-    // of the next file's (which is just younger) oldest entry's timestamp.
-    FileMetaData* prev_file = nullptr;
-    for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
-      FileMetaData* f = *ritr;
-      assert(f);
-      if (f->being_compacted) {
-        // Right now this probably won't happen as we never try to schedule
-        // two compactions in parallel, so here we just simply don't schedule
-        // anything.
+    // We will ideally identify a file qualifying for temperature change by
+    // knowing the timestamp for the youngest entry in the file. However, right
+    // now we don't have the information. We infer it by looking at timestamp of
+    // the previous file's (which is just younger) oldest entry's timestamp.
+    Temperature cur_target_temp;
+    // avoid index underflow
+    assert(level_files.size() >= 1);
+    for (size_t index = level_files.size() - 1; index >= 1; --index) {
+      // Try to add cur_file to compaction inputs.
+      FileMetaData* cur_file = level_files[index];
+      // prev_file is just younger than cur_file
+      FileMetaData* prev_file = level_files[index - 1];
+      if (cur_file->being_compacted) {
+        // Should not happen since we check for
+        // `level0_compactions_in_progress_` above. Here we simply just don't
+        // schedule anything.
        return nullptr;
      }
-      uint64_t oldest_ancester_time = f->TryGetOldestAncesterTime();
-      if (oldest_ancester_time == kUnknownOldestAncesterTime) {
+      uint64_t oldest_ancestor_time = prev_file->TryGetOldestAncesterTime();
+      if (oldest_ancestor_time == kUnknownOldestAncesterTime) {
        // Older files might not have enough information. It is possible to
        // handle these files by looking at newer files, but maintaining the
        // logic isn't worth it.
        break;
      }
-      if (oldest_ancester_time > create_time_threshold) {
-        // The previous file (which has slightly older data) doesn't qualify
-        // for warm tier.
+      if (oldest_ancestor_time > create_time_threshold) {
+        // cur_file is too fresh
        break;
      }
-      if (prev_file != nullptr) {
-        compaction_size += prev_file->fd.GetFileSize();
-        if (compaction_size > mutable_cf_options.max_compaction_bytes) {
+      cur_target_temp = ages[0].temperature;
+      for (size_t i = 1; i < ages.size(); ++i) {
+        if (current_time >= ages[i].age &&
+            oldest_ancestor_time <= current_time - ages[i].age) {
+          cur_target_temp = ages[i].temperature;
+        }
+      }
+      if (cur_file->temperature == cur_target_temp) {
+        if (inputs[0].empty()) {
+          continue;
+        } else {
          break;
        }
-        inputs[0].files.push_back(prev_file);
-        ROCKS_LOG_BUFFER(log_buffer,
-                         "[%s] FIFO compaction: picking file %" PRIu64
-                         " with next file's oldest time %" PRIu64 " for warm",
-                         cf_name.c_str(), prev_file->fd.GetNumber(),
-                         oldest_ancester_time);
      }
-      if (f->temperature == Temperature::kUnknown ||
-          f->temperature == Temperature::kHot) {
-        prev_file = f;
-      } else if (!inputs[0].files.empty()) {
-        // A warm file newer than files picked.
+
+      // cur_file needs to change temperature
+      if (compaction_target_temp == Temperature::kLastTemperature) {
+        assert(inputs[0].empty());
+        compaction_target_temp = cur_target_temp;
+      } else if (cur_target_temp != compaction_target_temp) {
+        assert(!inputs[0].empty());
+        break;
+      }
+      if (inputs[0].empty() || compaction_size + cur_file->fd.GetFileSize() <=
+                                   mutable_cf_options.max_compaction_bytes) {
+        inputs[0].files.push_back(cur_file);
+        compaction_size += cur_file->fd.GetFileSize();
+        ROCKS_LOG_BUFFER(
+            log_buffer,
+            "[%s] FIFO compaction: picking file %" PRIu64
+            " with next file's oldest time %" PRIu64 " for temperature %s.",
+            cf_name.c_str(), cur_file->fd.GetNumber(), oldest_ancestor_time,
+            temperature_to_string[cur_target_temp].c_str());
+      }
+      if (compaction_size > mutable_cf_options.max_compaction_bytes) {
        break;
-      } else {
-        assert(prev_file == nullptr);
      }
    }
  }
@@ -390,7 +417,7 @@ Compaction* FIFOCompactionPicker::PickCompactionToWarm(
      std::move(inputs), 0, 0 /* output file size limit */,
      0 /* max compaction bytes, not applicable */, 0 /* output path ID */,
      mutable_cf_options.compression, mutable_cf_options.compression_opts,
-      Temperature::kWarm,
+      compaction_target_temp,
      /* max_subcompactions */ 0, {}, /* is manual */ false, /* trim_ts */ "",
      vstorage->CompactionScore(0),
      /* is deletion compaction */ false, /* l0_files_might_overlap */ true,
@@ -412,8 +439,8 @@ Compaction* FIFOCompactionPicker::PickCompaction(
                           vstorage, log_buffer);
  }
  if (c == nullptr) {
-    c = PickCompactionToWarm(cf_name, mutable_cf_options, mutable_db_options,
-                             vstorage, log_buffer);
+    c = PickTemperatureChangeCompaction(
+        cf_name, mutable_cf_options, mutable_db_options, vstorage, log_buffer);
  }
  RegisterCompaction(c);
  return c;
@@ -52,10 +52,9 @@ class FIFOCompactionPicker : public CompactionPicker {
                                 VersionStorageInfo* version,
                                 LogBuffer* log_buffer);

-  Compaction* PickCompactionToWarm(const std::string& cf_name,
-                                   const MutableCFOptions& mutable_cf_options,
-                                   const MutableDBOptions& mutable_db_options,
-                                   VersionStorageInfo* version,
-                                   LogBuffer* log_buffer);
+  Compaction* PickTemperatureChangeCompaction(
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      const MutableDBOptions& mutable_db_options, VersionStorageInfo* vstorage,
+      LogBuffer* log_buffer);
 };
 }  // namespace ROCKSDB_NAMESPACE
@@ -148,7 +148,7 @@ class CompactionPickerTestBase : public testing::Test {
        smallest_seq, largest_seq, marked_for_compact, temperature,
        kInvalidBlobFileNumber, kUnknownOldestAncesterTime,
        kUnknownFileCreationTime, epoch_number, kUnknownFileChecksum,
-        kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+        kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
    f->compensated_file_size =
        (compensated_file_size != 0) ? compensated_file_size : file_size;
    f->oldest_ancester_time = oldest_ancestor_time;
@@ -1005,29 +1005,28 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
  }
 }

-TEST_F(CompactionPickerTest, FIFOToWarm1) {
+TEST_F(CompactionPickerTest, FIFOToCold1) {
  NewVersionStorage(1, kCompactionStyleFIFO);
  const uint64_t kFileSize = 100000;
  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
+  uint64_t kColdThreshold = 2000;

  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kCold, kColdThreshold}};
  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);

  int64_t current_time = 0;
  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
-  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
-      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
-  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
-      Temperature::kUnknown, threshold_time + 100);
-  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
-      Temperature::kUnknown, threshold_time - 2000);
+      static_cast<uint64_t>(current_time) - kColdThreshold;
+  Add(0 /* level */, 4U /* file_number */, "260", "300", 1 * kFileSize, 0, 2500,
+      2600, 0, true, Temperature::kUnknown,
+      threshold_time - 2000 /* oldest_ancestor_time */);
+  // Qualifies for compaction to kCold.
  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
      Temperature::kUnknown, threshold_time - 3000);
  UpdateVersionStorageInfo();
@@ -1037,33 +1036,36 @@ TEST_F(CompactionPickerTest, FIFOToWarm1) {
      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
      &log_buffer_));
  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kCold);
  ASSERT_EQ(1U, compaction->num_input_files(0));
  ASSERT_EQ(3U, compaction->input(0, 0)->fd.GetNumber());
 }

-TEST_F(CompactionPickerTest, FIFOToWarm2) {
+TEST_F(CompactionPickerTest, FIFOToCold2) {
  NewVersionStorage(1, kCompactionStyleFIFO);
  const uint64_t kFileSize = 100000;
  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
+  uint64_t kColdThreshold = 2000;

  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kCold, kColdThreshold}};
  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);

  int64_t current_time = 0;
  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
+      static_cast<uint64_t>(current_time) - kColdThreshold;
  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
-  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
-      Temperature::kUnknown, threshold_time + 100);
  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
-      Temperature::kUnknown, threshold_time - 2000);
+      Temperature::kUnknown, threshold_time);
+  // The following two files qualify for compaction to kCold.
  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
      Temperature::kUnknown, threshold_time - 3000);
  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
@@ -1075,34 +1077,40 @@ TEST_F(CompactionPickerTest, FIFOToWarm2) {
      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
      &log_buffer_));
  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kCold);
  ASSERT_EQ(2U, compaction->num_input_files(0));
  ASSERT_EQ(2U, compaction->input(0, 0)->fd.GetNumber());
  ASSERT_EQ(3U, compaction->input(0, 1)->fd.GetNumber());
 }

-TEST_F(CompactionPickerTest, FIFOToWarmMaxSize) {
+TEST_F(CompactionPickerTest, FIFOToColdMaxCompactionSize) {
  NewVersionStorage(1, kCompactionStyleFIFO);
  const uint64_t kFileSize = 100000;
  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
+  uint64_t kColdThreshold = 2000;

  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kCold, kColdThreshold}};
  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
  mutable_cf_options_.max_compaction_bytes = kFileSize * 9;
  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);

  int64_t current_time = 0;
  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
+      static_cast<uint64_t>(current_time) - kColdThreshold;
  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
      Temperature::kUnknown, threshold_time + 100);
  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
      Temperature::kUnknown, threshold_time - 2000);
+  // The following two files qualify for compaction to kCold.
+  // But only the last two should be included to respect `max_compaction_bytes`.
  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
      Temperature::kUnknown, threshold_time - 3000);
  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
@@ -1116,40 +1124,45 @@ TEST_F(CompactionPickerTest, FIFOToWarmMaxSize) {
      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
      &log_buffer_));
  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kCold);
  ASSERT_EQ(2U, compaction->num_input_files(0));
  ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
  ASSERT_EQ(2U, compaction->input(0, 1)->fd.GetNumber());
 }

-TEST_F(CompactionPickerTest, FIFOToWarmWithExistingWarm) {
+TEST_F(CompactionPickerTest, FIFOToColdWithExistingCold) {
  NewVersionStorage(1, kCompactionStyleFIFO);
  const uint64_t kFileSize = 100000;
  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
+  uint64_t kColdThreshold = 2000;

  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kCold, kColdThreshold}};
  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);

  int64_t current_time = 0;
  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
+      static_cast<uint64_t>(current_time) - kColdThreshold;
  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
      Temperature::kUnknown, threshold_time + 100);
  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
      Temperature::kUnknown, threshold_time - 2000);
+  // The following two files qualify for compaction to kCold.
  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
      Temperature::kUnknown, threshold_time - 3000);
  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
      Temperature::kUnknown, threshold_time - 4000);
  Add(0, 1U, "200", "300", 4 * kFileSize, 0, 2000, 2100, 0, true,
-      Temperature::kWarm, threshold_time - 5000);
+      Temperature::kCold, threshold_time - 5000);
  UpdateVersionStorageInfo();

  ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()), true);
@@ -1157,28 +1170,32 @@ TEST_F(CompactionPickerTest, FIFOToWarmWithExistingWarm) {
      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
      &log_buffer_));
  ASSERT_TRUE(compaction.get() != nullptr);
-  ASSERT_EQ(2U, compaction->num_input_files(0));
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kCold);
  ASSERT_EQ(2U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(2U, compaction->num_input_files(0));
  ASSERT_EQ(3U, compaction->input(0, 1)->fd.GetNumber());
 }

-TEST_F(CompactionPickerTest, FIFOToWarmWithOngoing) {
+TEST_F(CompactionPickerTest, FIFOToColdWithHotBetweenCold) {
  NewVersionStorage(1, kCompactionStyleFIFO);
  const uint64_t kFileSize = 100000;
  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
+  uint64_t kColdThreshold = 2000;

  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kCold, kColdThreshold}};
  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);

  int64_t current_time = 0;
  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
+      static_cast<uint64_t>(current_time) - kColdThreshold;
  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
@@ -1186,63 +1203,77 @@ TEST_F(CompactionPickerTest, FIFOToWarmWithOngoing) {
  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
      Temperature::kUnknown, threshold_time - 2000);
  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
-      Temperature::kUnknown, threshold_time - 3000);
+      Temperature::kCold, threshold_time - 3000);
+  // Qualifies for compaction to kCold.
  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
      Temperature::kUnknown, threshold_time - 4000);
  Add(0, 1U, "200", "300", 4 * kFileSize, 0, 2000, 2100, 0, true,
-      Temperature::kWarm, threshold_time - 5000);
-  file_map_[2].first->being_compacted = true;
+      Temperature::kCold, threshold_time - 5000);
  UpdateVersionStorageInfo();

  ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()), true);
  std::unique_ptr<Compaction> compaction(fifo_compaction_picker.PickCompaction(
      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
      &log_buffer_));
-  // Stop if a file is being compacted
-  ASSERT_TRUE(compaction.get() == nullptr);
-}
-
-TEST_F(CompactionPickerTest, FIFOToWarmWithHotBetweenWarms) {
-  NewVersionStorage(1, kCompactionStyleFIFO);
-  const uint64_t kFileSize = 100000;
-  const uint64_t kMaxSize = kFileSize * 100000;
-  uint64_t kWarmThreshold = 2000;
-
-  fifo_options_.max_table_files_size = kMaxSize;
-  fifo_options_.age_for_warm = kWarmThreshold;
-  mutable_cf_options_.compaction_options_fifo = fifo_options_;
-  mutable_cf_options_.level0_file_num_compaction_trigger = 2;
-  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
-  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);
-
-  int64_t current_time = 0;
-  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
-  uint64_t threshold_time =
-      static_cast<uint64_t>(current_time) - kWarmThreshold;
-  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
-      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
-  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
-      Temperature::kUnknown, threshold_time + 100);
-  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
-      Temperature::kUnknown, threshold_time - 2000);
-  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
-      Temperature::kWarm, threshold_time - 3000);
-  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
-      Temperature::kUnknown, threshold_time - 4000);
-  Add(0, 1U, "200", "300", 4 * kFileSize, 0, 2000, 2100, 0, true,
-      Temperature::kWarm, threshold_time - 5000);
-  UpdateVersionStorageInfo();
-
-  ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()), true);
-  std::unique_ptr<Compaction> compaction(fifo_compaction_picker.PickCompaction(
-      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
-      &log_buffer_));
-  // Stop if a file is being compacted
  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kCold);
  ASSERT_EQ(1U, compaction->num_input_files(0));
  ASSERT_EQ(2U, compaction->input(0, 0)->fd.GetNumber());
 }

+TEST_F(CompactionPickerTest, FIFOToColdAndWarm) {
+  NewVersionStorage(1, kCompactionStyleFIFO);
+  const uint64_t kFileSize = 100000;
+  const uint64_t kMaxSize = kFileSize * 100000;
+  uint64_t kWarmThreshold = 10000;
+  uint64_t kHotThreshold = 2000;
+
+  fifo_options_.max_table_files_size = kMaxSize;
+  // Test that multiple threshold works.
+  fifo_options_.file_temperature_age_thresholds = {
+      {Temperature::kHot, kHotThreshold}, {Temperature::kWarm, kWarmThreshold}};
+  mutable_cf_options_.compaction_options_fifo = fifo_options_;
+  mutable_cf_options_.level0_file_num_compaction_trigger = 100;
+  mutable_cf_options_.max_compaction_bytes = kFileSize * 100;
+  FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);
+
+  int64_t current_time = 0;
+  ASSERT_OK(Env::Default()->GetCurrentTime(&current_time));
+  uint64_t hot_threshold_time =
+      static_cast<uint64_t>(current_time) - kHotThreshold;
+  uint64_t warm_threshold_time =
+      static_cast<uint64_t>(current_time) - kWarmThreshold;
+  Add(0, 6U, "240", "290", 2 * kFileSize, 0, 2900, 3000, 0, true,
+      Temperature::kUnknown, static_cast<uint64_t>(current_time) - 100);
+  Add(0, 5U, "240", "290", 2 * kFileSize, 0, 2700, 2800, 0, true,
+      Temperature::kUnknown, hot_threshold_time + 100);
+  Add(0, 4U, "260", "300", 1 * kFileSize, 0, 2500, 2600, 0, true,
+      Temperature::kUnknown, hot_threshold_time - 200);
+  // Qualifies for Hot
+  Add(0, 3U, "200", "300", 4 * kFileSize, 0, 2300, 2400, 0, true,
+      Temperature::kUnknown, warm_threshold_time - 100);
+  // Qualifies for Warm
+  Add(0, 2U, "200", "300", 4 * kFileSize, 0, 2100, 2200, 0, true,
+      Temperature::kUnknown, warm_threshold_time - 4000);
+  Add(0, 1U, "200", "300", 4 * kFileSize, 0, 2000, 2100, 0, true,
+      Temperature::kUnknown, warm_threshold_time - 5000);
+  UpdateVersionStorageInfo();
+
+  ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()), true);
+  std::unique_ptr<Compaction> compaction(fifo_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, mutable_db_options_, vstorage_.get(),
+      &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(compaction->compaction_reason(),
+            CompactionReason::kChangeTemperature);
+  // Assumes compaction picker picks older files first.
+  ASSERT_EQ(compaction->output_temperature(), Temperature::kWarm);
+  ASSERT_EQ(2U, compaction->num_input_files(0));
+  ASSERT_EQ(1U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(2U, compaction->input(0, 1)->fd.GetNumber());
+}

 TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
  NewVersionStorage(6, kCompactionStyleLevel);
@@ -19,7 +19,7 @@
 #include "file/filename.h"
 #include "logging/log_buffer.h"
 #include "logging/logging.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "test_util/sync_point.h"
 #include "util/random.h"
 #include "util/string_util.h"
@@ -928,7 +928,7 @@ TEST_F(CompactionServiceTest, TablePropertiesCollector) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_OK(db_->GetPropertiesOfAllTables(&fname_to_props));

@@ -952,4 +952,3 @@ int main(int argc, char** argv) {
  RegisterCustomObjects(argc, argv);
  return RUN_ALL_TESTS();
 }
-
@@ -209,7 +209,7 @@ TEST_P(TieredCompactionTest, SequenceBasedTieredStorageUniversal) {
    seq_history.emplace_back(dbfull()->GetLatestSequenceNumber());
    expect_stats[0].Add(kBasicFlushStats);
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // the penultimate level file temperature is not cold, all data are output to
  // the penultimate level.
@@ -374,7 +374,7 @@ TEST_P(TieredCompactionTest, RangeBasedTieredStorageUniversal) {
    ASSERT_OK(Flush());
    expect_stats[0].Add(kBasicFlushStats);
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,1,1", FilesPerLevel());
  ASSERT_GT(GetSstSizeHelper(Temperature::kUnknown), 0);
  ASSERT_GT(GetSstSizeHelper(Temperature::kCold), 0);
@@ -445,8 +445,8 @@ TEST_P(TieredCompactionTest, RangeBasedTieredStorageUniversal) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(
-      true));  // make sure the compaction is able to finish
+  // make sure the compaction is able to finish
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,1,1", FilesPerLevel());
  ASSERT_GT(GetSstSizeHelper(Temperature::kUnknown), 0);
  ASSERT_GT(GetSstSizeHelper(Temperature::kCold), 0);
@@ -911,7 +911,7 @@ TEST_P(TieredCompactionTest, SequenceBasedTieredStorageLevel) {
    ASSERT_OK(Flush());
    expect_stats[0].Add(kBasicFlushStats);
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // non last level is hot
  ASSERT_EQ("0,1", FilesPerLevel());
@@ -954,7 +954,7 @@ TEST_P(TieredCompactionTest, SequenceBasedTieredStorageLevel) {
    ASSERT_OK(Flush());
    seq_history.emplace_back(dbfull()->GetLatestSequenceNumber());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,1,0,0,0,0,1", FilesPerLevel());
  ASSERT_GT(GetSstSizeHelper(Temperature::kUnknown), 0);
  ASSERT_GT(GetSstSizeHelper(Temperature::kCold), 0);
@@ -1005,7 +1005,7 @@ TEST_P(TieredCompactionTest, SequenceBasedTieredStorageLevel) {
    ASSERT_OK(Flush());
    seq_history.emplace_back(dbfull()->GetLatestSequenceNumber());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  latest_cold_seq = seq_history[0];
  ASSERT_OK(db_->CompactRange(cro, nullptr, nullptr));
@@ -1135,7 +1135,7 @@ TEST_P(TieredCompactionTest, RangeBasedTieredStorageLevel) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,1,1", FilesPerLevel());
  ASSERT_GT(GetSstSizeHelper(Temperature::kUnknown), 0);
  ASSERT_GT(GetSstSizeHelper(Temperature::kCold), 0);
@@ -1265,7 +1265,7 @@ TEST_F(PrecludeLastLevelTest, MigrationFromPreserveTimeManualCompaction) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // all data is pushed to the last level
  ASSERT_EQ("0,0,0,0,0,0,1", FilesPerLevel());
@@ -1327,7 +1327,7 @@ TEST_F(PrecludeLastLevelTest, MigrationFromPreserveTimeAutoCompaction) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // all data is pushed to the last level
  ASSERT_EQ("0,0,0,0,0,0,1", FilesPerLevel());
@@ -1360,7 +1360,7 @@ TEST_F(PrecludeLastLevelTest, MigrationFromPreserveTimeAutoCompaction) {
      });
    }
    ASSERT_OK(Flush());
-    ASSERT_OK(dbfull()->WaitForCompact(true));
+    ASSERT_OK(dbfull()->WaitForCompact());
  }

  // all data is moved up to the penultimate level
@@ -1403,7 +1403,7 @@ TEST_F(PrecludeLastLevelTest, MigrationFromPreserveTimePartial) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // all data is pushed to the last level
  ASSERT_EQ("0,0,0,0,0,0,1", FilesPerLevel());
@@ -1530,7 +1530,7 @@ TEST_F(PrecludeLastLevelTest, LastLevelOnlyCompactionPartial) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // all data is pushed to the last level
  ASSERT_EQ("0,0,0,0,0,0,1", FilesPerLevel());
@@ -1609,7 +1609,7 @@ TEST_P(PrecludeLastLevelTestWithParms, LastLevelOnlyCompactionNoPreclude) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // all data is pushed to the last level
  ASSERT_EQ("0,0,0,0,0,0,1", FilesPerLevel());
@@ -1705,7 +1705,7 @@ TEST_P(PrecludeLastLevelTestWithParms, LastLevelOnlyCompactionNoPreclude) {

  manual_compaction_thread.join();

-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  if (enable_preclude_last_level) {
    ASSERT_NE("0,0,0,0,0,1,1", FilesPerLevel());
@@ -1841,7 +1841,7 @@ TEST_P(PrecludeLastLevelTestWithParms, PeriodicCompactionToPenultimateLevel) {
  }
  ASSERT_OK(Flush());

-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  stop_token.reset();

@@ -1940,7 +1940,7 @@ TEST_F(PrecludeLastLevelTest, PartialPenultimateLevelCompaction) {
    ASSERT_OK(Flush());
  }

-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // L5: [0,19] [20,39] [40,299]
  // L6: [0,                299]
@@ -2106,7 +2106,7 @@ TEST_F(PrecludeLastLevelTest, RangeDelsCauseFileEndpointsToOverlap) {
  Slice begin_key(begin_key_buf), end_key(end_key_buf);
  ASSERT_OK(db_->SuggestCompactRange(db_->DefaultColumnFamily(), &begin_key,
                                     &end_key));
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,3,3", FilesPerLevel());
  ASSERT_EQ(1, per_key_comp_num);
  verify_db();
@@ -2116,7 +2116,7 @@ TEST_F(PrecludeLastLevelTest, RangeDelsCauseFileEndpointsToOverlap) {
  db_->ReleaseSnapshot(snap2);
  ASSERT_OK(db_->SuggestCompactRange(db_->DefaultColumnFamily(), &begin_key,
                                     &end_key));
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,3,3", FilesPerLevel());
  ASSERT_EQ(2, per_key_comp_num);
  verify_db();
@@ -2126,7 +2126,7 @@ TEST_F(PrecludeLastLevelTest, RangeDelsCauseFileEndpointsToOverlap) {
  db_->ReleaseSnapshot(snap1);
  ASSERT_OK(db_->SuggestCompactRange(db_->DefaultColumnFamily(), &begin_key,
                                     &end_key));
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,2,3", FilesPerLevel());
  ASSERT_EQ(3, per_key_comp_num);
  verify_db();
@@ -64,8 +64,8 @@ Status VerifySstFileChecksum(const Options& options,
  const bool kImmortal = true;
  auto reader_options = TableReaderOptions(
      ioptions, options.prefix_extractor, env_options, internal_comparator,
-      false /* skip_filters */, !kImmortal, false /* force_direct_prefetch */,
-      -1 /* level */);
+      options.block_protection_bytes_per_key, false /* skip_filters */,
+      !kImmortal, false /* force_direct_prefetch */, -1 /* level */);
  reader_options.largest_seqno = largest_seqno;
  s = ioptions.table_factory->NewTableReader(
      reader_options, std::move(file_reader), file_size, &table_reader,
@@ -620,9 +620,9 @@ class MockCache : public LRUCache {
  static uint32_t low_pri_insert_count;

  MockCache()
-      : LRUCache((size_t)1 << 25 /*capacity*/, 0 /*num_shard_bits*/,
-                 false /*strict_capacity_limit*/, 0.0 /*high_pri_pool_ratio*/,
-                 0.0 /*low_pri_pool_ratio*/) {}
+      : LRUCache(LRUCacheOptions(
+            size_t{1} << 25 /*capacity*/, 0 /*num_shard_bits*/,
+            false /*strict_capacity_limit*/, 0.0 /*high_pri_pool_ratio*/)) {}

  using ShardedCache::Insert;

@@ -979,12 +979,14 @@ TEST_F(DBBlockCacheTest, CacheEntryRoleStats) {
  const size_t capacity = size_t{1} << 25;
  int iterations_tested = 0;
  for (bool partition : {false, true}) {
+    SCOPED_TRACE("Partition? " + std::to_string(partition));
    for (std::shared_ptr<Cache> cache :
         {NewLRUCache(capacity),
          HyperClockCacheOptions(
              capacity,
              BlockBasedTableOptions().block_size /*estimated_value_size*/)
              .MakeSharedCache()}) {
+      SCOPED_TRACE(std::string("Cache: ") + cache->Name());
      ++iterations_tested;

      Options options = CurrentOptions();
@@ -1278,6 +1280,7 @@ TEST_F(DBBlockCacheTest, HyperClockCacheReportProblems) {
  HyperClockCacheOptions hcc_opts{capacity, value_size_est};
  hcc_opts.num_shard_bits = 2;  // 4 shards
  hcc_opts.metadata_charge_policy = kDontChargeCacheMetadata;
+  hcc_opts.hash_seed = 0;  // deterministic hashing
  std::shared_ptr<Cache> cache = hcc_opts.MakeSharedCache();
  std::shared_ptr<CountingLogger> logger = std::make_shared<CountingLogger>();

@@ -43,6 +43,16 @@ const std::string kStandard128Ribbon =
    test::Standard128RibbonFilterPolicy::kClassName();
 const std::string kAutoBloom = BloomFilterPolicy::kClassName();
 const std::string kAutoRibbon = RibbonFilterPolicy::kClassName();
+
+template <typename T>
+T Pop(T& var) {
+  auto rv = var;
+  var = 0;
+  return rv;
+}
+PerfContextByLevel& GetLevelPerfContext(uint32_t level) {
+  return (*(get_perf_context()->level_to_perf_context))[level];
+}
 }  // anonymous namespace

 // DB tests related to bloom filter.
@@ -209,57 +219,43 @@ TEST_F(DBBloomFilterTest, GetFilterByPrefixBloomCustomPrefixExtractor) {
    ASSERT_OK(dbfull()->Flush(fo));

    ASSERT_EQ("foo", Get("barbarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-    ASSERT_EQ(
-        0,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);
+
    ASSERT_EQ("foo2", Get("barbarbar2"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-    ASSERT_EQ(
-        0,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);
+
    ASSERT_EQ("NOT_FOUND", Get("barbarbar3"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
-    ASSERT_EQ(
-        0,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);

    ASSERT_EQ("NOT_FOUND", Get("barfoofoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
-    ASSERT_EQ(
-        1,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
-    ASSERT_EQ(
-        2,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    ro.total_order_seek = true;
    // NOTE: total_order_seek no longer affects Get()
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-    ASSERT_EQ(
-        3,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    // No bloom on extractor changed
    ASSERT_OK(db_->SetOptions({{"prefix_extractor", "capped:10"}}));
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-    ASSERT_EQ(
-        3,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);

    // No bloom on extractor changed, after re-open
    options.prefix_extractor.reset(NewCappedPrefixTransform(10));
    Reopen(options);
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-    ASSERT_EQ(
-        3,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);

    get_perf_context()->Reset();
  }
@@ -294,33 +290,32 @@ TEST_F(DBBloomFilterTest, GetFilterByPrefixBloom) {
    ASSERT_OK(dbfull()->Flush(fo));

    ASSERT_EQ("foo", Get("barbarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
    ASSERT_EQ("foo2", Get("barbarbar2"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("barbarbar3"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);

    ASSERT_EQ("NOT_FOUND", Get("barfoofoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    ro.total_order_seek = true;
    // NOTE: total_order_seek no longer affects Get()
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-    ASSERT_EQ(
-        3,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 1);

    // No bloom on extractor changed
    ASSERT_OK(db_->SetOptions({{"prefix_extractor", "capped:10"}}));
    ASSERT_EQ("NOT_FOUND", Get("foobarbar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
-    ASSERT_EQ(
-        3,
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(Pop(GetLevelPerfContext(0).bloom_filter_useful), 0);

    get_perf_context()->Reset();
  }
@@ -357,12 +352,17 @@ TEST_F(DBBloomFilterTest, WholeKeyFilterProp) {
    ASSERT_OK(dbfull()->Flush(fo));

    Reopen(options);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);

    // Reopen with whole key filtering enabled and prefix extractor
    // NULL. Bloom filter should be off for both of whole key and
@@ -372,13 +372,17 @@ TEST_F(DBBloomFilterTest, WholeKeyFilterProp) {
    options.prefix_extractor.reset();
    Reopen(options);

-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    // Write DB with only full key filtering.
    ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
    // Needs insert some keys to make sure files are not filtered out by key
@@ -394,13 +398,17 @@ TEST_F(DBBloomFilterTest, WholeKeyFilterProp) {
    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
    Reopen(options);

-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);

    // Try to create a DB with mixed files:
    ASSERT_OK(dbfull()->Put(wo, "foobar", "foo"));
@@ -424,61 +432,81 @@ TEST_F(DBBloomFilterTest, WholeKeyFilterProp) {
    ASSERT_OK(Flush());

    // Now we have two files:
-    // File 1: An older file with prefix bloom.
+    // File 1: An older file with prefix bloom (disabled)
    // File 2: A newer file with whole bloom filter.
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 2);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 3);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("bar", Get("barfoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);

    // Reopen with the same setting: only whole key is used
    Reopen(options);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 4);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 5);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 6);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("bar", Get("barfoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);

    // Restart with both filters are allowed
    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
    bbto.whole_key_filtering = true;
    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
    Reopen(options);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 7);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    // File 1 will has it filtered out.
    // File 2 will not, as prefix `foo` exists in the file.
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 8);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 10);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    ASSERT_EQ("bar", Get("barfoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);

    // Restart with only prefix bloom is allowed.
    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
    bbto.whole_key_filtering = false;
    options.table_factory.reset(NewBlockBasedTableFactory(bbto));
    Reopen(options);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("foo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 11);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("NOT_FOUND", Get("bar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("foo", Get("foobar"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    ASSERT_EQ("bar", Get("barfoo"));
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), 12);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 0);
    uint64_t bloom_filter_useful_all_levels = 0;
    for (auto& kv : (*(get_perf_context()->level_to_perf_context))) {
      if (kv.second.bloom_filter_useful > 0) {
@@ -696,11 +724,8 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  PutAndGetFn();

  // Verify filter is accessed (and constructed)
-  EXPECT_EQ(TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE),
-            maxKey * 2);
-  EXPECT_EQ(
-      TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE),
-      maxKey);
+  EXPECT_EQ(PopTicker(options, BLOOM_FILTER_FULL_POSITIVE), maxKey * 2);
+  EXPECT_EQ(PopTicker(options, BLOOM_FILTER_FULL_TRUE_POSITIVE), maxKey);
  props.clear();
  ASSERT_TRUE(db_->GetMapProperty(kAggTableProps, &props));
  EXPECT_NE(props["filter_size"], "0");
@@ -715,11 +740,8 @@ TEST_P(DBBloomFilterTestWithParam, SkipFilterOnEssentiallyZeroBpk) {
  GetFn();

  // Verify filter is accessed
-  EXPECT_EQ(TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE),
-            maxKey * 2);
-  EXPECT_EQ(
-      TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE),
-      maxKey);
+  EXPECT_EQ(PopTicker(options, BLOOM_FILTER_FULL_POSITIVE), maxKey * 2);
+  EXPECT_EQ(PopTicker(options, BLOOM_FILTER_FULL_TRUE_POSITIVE), maxKey);

  // But new filters are not generated (configuration details unknown)
  DestroyAndReopen(options);
@@ -783,9 +805,7 @@ TEST_F(DBBloomFilterTest, BloomFilterRate) {
      ASSERT_EQ("NOT_FOUND", Get(1, Key(i + 33333)));
    }
    ASSERT_GE(TestGetTickerCount(options, BLOOM_FILTER_USEFUL), maxKey * 0.98);
-    ASSERT_GE(
-        (*(get_perf_context()->level_to_perf_context))[0].bloom_filter_useful,
-        maxKey * 0.98);
+    ASSERT_GE(GetLevelPerfContext(0).bloom_filter_useful, maxKey * 0.98);
    get_perf_context()->Reset();
  }
 }
@@ -864,9 +884,8 @@ TEST_F(DBBloomFilterTest, BloomFilterCompatibility) {
      ASSERT_EQ("val", Get(prefix + "Z"));  // Filter positive
      // Filter negative, with high probability
      ASSERT_EQ("NOT_FOUND", Get(prefix + "Q"));
-      EXPECT_EQ(TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE),
-                2);
-      EXPECT_EQ(TestGetAndResetTickerCount(options, BLOOM_FILTER_USEFUL), 1);
+      EXPECT_EQ(PopTicker(options, BLOOM_FILTER_FULL_POSITIVE), 2);
+      EXPECT_EQ(PopTicker(options, BLOOM_FILTER_USEFUL), 1);
    }
  }
 }
@@ -1705,7 +1724,7 @@ TEST_F(DBBloomFilterTest, ContextCustomFilterPolicy) {
    }
    // Since we have two tables / two filters, we might have Bloom checks on
    // our queries, but no more than one "useful" per query on a found key.
-    EXPECT_LE(TestGetAndResetTickerCount(options, BLOOM_FILTER_USEFUL), maxKey);
+    EXPECT_LE(PopTicker(options, BLOOM_FILTER_USEFUL), maxKey);

    // Check that we have two filters, each about
    // fifo: 0.12% FP rate (15 bits per key)
@@ -1714,8 +1733,7 @@ TEST_F(DBBloomFilterTest, ContextCustomFilterPolicy) {
      ASSERT_EQ("NOT_FOUND", Get(1, Key(i + 33333)));
    }
    {
-      auto useful_count =
-          TestGetAndResetTickerCount(options, BLOOM_FILTER_USEFUL);
+      auto useful_count = PopTicker(options, BLOOM_FILTER_USEFUL);
      EXPECT_GE(useful_count, maxKey * 2 * (fifo ? 0.9980 : 0.975));
      EXPECT_LE(useful_count, maxKey * 2 * (fifo ? 0.9995 : 0.98));
    }
@@ -1732,8 +1750,7 @@ TEST_F(DBBloomFilterTest, ContextCustomFilterPolicy) {
        ASSERT_EQ("NOT_FOUND", Get(1, Key(i + 33333)));
      }
      {
-        auto useful_count =
-            TestGetAndResetTickerCount(options, BLOOM_FILTER_USEFUL);
+        auto useful_count = PopTicker(options, BLOOM_FILTER_USEFUL);
        EXPECT_GE(useful_count, maxKey * 0.90);
        EXPECT_LE(useful_count, maxKey * 0.91);
      }
@@ -2027,13 +2044,14 @@ TEST_P(DBBloomFilterTestVaryPrefixAndFormatVer, PartitionedMultiGet) {
  std::array<Status, Q> statuses;
  std::array<PinnableSlice, Q> values;

-  TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_HIT);
-  TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_MISS);
-  TestGetAndResetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL);
-  TestGetAndResetTickerCount(options, BLOOM_FILTER_USEFUL);
-  TestGetAndResetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED);
-  TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_POSITIVE);
-  TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE);
+  PopTicker(options, BLOCK_CACHE_FILTER_HIT);
+  PopTicker(options, BLOCK_CACHE_FILTER_MISS);
+  PopTicker(options, BLOOM_FILTER_PREFIX_USEFUL);
+  PopTicker(options, BLOOM_FILTER_USEFUL);
+  PopTicker(options, BLOOM_FILTER_PREFIX_CHECKED);
+  PopTicker(options, BLOOM_FILTER_FULL_POSITIVE);
+  PopTicker(options, BLOOM_FILTER_FULL_TRUE_POSITIVE);
+  PopTicker(options, BLOOM_FILTER_PREFIX_TRUE_POSITIVE);

  // Check that initial clump of keys only loads one partition filter from
  // block cache.
@@ -2063,26 +2081,22 @@ TEST_P(DBBloomFilterTestVaryPrefixAndFormatVer, PartitionedMultiGet) {
    }

    // Confirm correct Bloom stats (no FPs)
-    uint64_t filter_useful = TestGetAndResetTickerCount(
-        options,
-        use_prefix_ ? BLOOM_FILTER_PREFIX_USEFUL : BLOOM_FILTER_USEFUL);
+    uint64_t filter_useful =
+        PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_USEFUL
+                                       : BLOOM_FILTER_USEFUL);
    uint64_t filter_checked =
-        TestGetAndResetTickerCount(options, use_prefix_
-                                                ? BLOOM_FILTER_PREFIX_CHECKED
-                                                : BLOOM_FILTER_FULL_POSITIVE) +
+        PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_CHECKED
+                                       : BLOOM_FILTER_FULL_POSITIVE) +
        (use_prefix_ ? 0 : filter_useful);
    EXPECT_EQ(filter_useful, number_not_found);
    EXPECT_EQ(filter_checked, Q);
-    if (!use_prefix_) {
-      EXPECT_EQ(
-          TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE),
-          Q - number_not_found);
-    }
+    EXPECT_EQ(PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_TRUE_POSITIVE
+                                             : BLOOM_FILTER_FULL_TRUE_POSITIVE),
+              Q - number_not_found);

    // Confirm no duplicate loading same filter partition
-    uint64_t filter_accesses =
-        TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_HIT) +
-        TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_MISS);
+    uint64_t filter_accesses = PopTicker(options, BLOCK_CACHE_FILTER_HIT) +
+                               PopTicker(options, BLOCK_CACHE_FILTER_MISS);
    if (stride == 1) {
      EXPECT_EQ(filter_accesses, 1);
    } else {
@@ -2118,26 +2132,22 @@ TEST_P(DBBloomFilterTestVaryPrefixAndFormatVer, PartitionedMultiGet) {
    }

    // Confirm correct Bloom stats (might see some FPs)
-    uint64_t filter_useful = TestGetAndResetTickerCount(
-        options,
-        use_prefix_ ? BLOOM_FILTER_PREFIX_USEFUL : BLOOM_FILTER_USEFUL);
+    uint64_t filter_useful =
+        PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_USEFUL
+                                       : BLOOM_FILTER_USEFUL);
    uint64_t filter_checked =
-        TestGetAndResetTickerCount(options, use_prefix_
-                                                ? BLOOM_FILTER_PREFIX_CHECKED
-                                                : BLOOM_FILTER_FULL_POSITIVE) +
+        PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_CHECKED
+                                       : BLOOM_FILTER_FULL_POSITIVE) +
        (use_prefix_ ? 0 : filter_useful);
    EXPECT_GE(filter_useful, number_not_found - 2);  // possible FP
    EXPECT_EQ(filter_checked, Q);
-    if (!use_prefix_) {
-      EXPECT_EQ(
-          TestGetAndResetTickerCount(options, BLOOM_FILTER_FULL_TRUE_POSITIVE),
-          Q - number_not_found);
-    }
+    EXPECT_EQ(PopTicker(options, use_prefix_ ? BLOOM_FILTER_PREFIX_TRUE_POSITIVE
+                                             : BLOOM_FILTER_FULL_TRUE_POSITIVE),
+              Q - number_not_found);

    // Confirm no duplicate loading of same filter partition
-    uint64_t filter_accesses =
-        TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_HIT) +
-        TestGetAndResetTickerCount(options, BLOCK_CACHE_FILTER_MISS);
+    uint64_t filter_accesses = PopTicker(options, BLOCK_CACHE_FILTER_HIT) +
+                               PopTicker(options, BLOCK_CACHE_FILTER_MISS);
    if (filter_accesses == 2) {
      // Spanned across partitions.
      ++found_spanning;
@@ -2636,6 +2646,9 @@ int CountIter(std::unique_ptr<Iterator>& iter, const Slice& key) {
  int count = 0;
  for (iter->Seek(key); iter->Valid(); iter->Next()) {
    count++;
+    // Access key & value as if we were using them
+    (void)iter->key();
+    (void)iter->value();
  }
  EXPECT_OK(iter->status());
  return count;
@@ -2675,6 +2688,12 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abcd0000"), 4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                1);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      ASSERT_EQ(TestGetTickerCount(
+                    options, NON_LAST_LEVEL_SEEK_DATA_USEFUL_FILTER_MATCH),
+                1);
    }
    {
      Slice upper_bound("abcdzzzz");
@@ -2683,8 +2702,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abcd0000"), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 2);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                2);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "fixed:5"}}));
    ASSERT_EQ(dbfull()->GetOptions().prefix_extractor->AsString(),
@@ -2698,8 +2718,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abcdxx00"), 4);
      // should check bloom filter since upper bound meets requirement
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 3);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                3);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    {
      // [abcdxx01, abcey) is not valid bound since upper bound is too long for
@@ -2711,8 +2732,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abcdxx01"), 4);
      // should skip bloom filter since upper bound is too long
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 3);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                3);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    {
      // [abcdxx02, abcdy) is a valid bound since the prefix is the same
@@ -2724,8 +2746,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      ASSERT_EQ(CountIter(iter, "abcdxx02"), 4);
      // should check bloom filter since upper bound matches transformed seek
      // key
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    {
      // [aaaaaaaa, abce) is not a valid bound since 1) they don't share the
@@ -2737,8 +2760,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "aaaaaaaa"), 0);
      // should skip bloom filter since mismatch is found
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "fixed:3"}}));
    {
@@ -2750,8 +2774,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abc"), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    // Same with re-open
    options.prefix_extractor.reset(NewFixedPrefixTransform(3));
@@ -2763,8 +2788,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abc"), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
    }
    // Set back to capped:4 and verify BF is always read
    options.prefix_extractor.reset(NewCappedPrefixTransform(4));
@@ -2776,8 +2802,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abc"), 0);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 5);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 1);
    }
    // Same if there's a problem initally loading prefix transform
    SyncPoint::GetInstance()->SetCallBack(
@@ -2792,8 +2819,9 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterUpperBound) {
      read_options.iterate_upper_bound = &upper_bound;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "abc"), 0);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 6);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 2);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+                4);
+      ASSERT_EQ(TestGetTickerCount(options, NON_LAST_LEVEL_SEEK_FILTERED), 2);
    }
    SyncPoint::GetInstance()->DisableProcessing();
  }
@@ -2828,7 +2856,8 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
    dbfull()->Flush(FlushOptions());
    std::unique_ptr<Iterator> iter_old(db_->NewIterator(read_options));
    ASSERT_EQ(CountIter(iter_old, "foo"), 4);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 1);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 1);

    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "capped:3"}}));
    ASSERT_EQ(dbfull()->GetOptions().prefix_extractor->AsString(),
@@ -2836,10 +2865,11 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
    read_options.iterate_upper_bound = &upper_bound;
    std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
    ASSERT_EQ(CountIter(iter, "foo"), 2);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 2);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 1);
    ASSERT_EQ(CountIter(iter, "gpk"), 0);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 2);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);

    // second SST with capped:3 BF
    ASSERT_OK(Put("foo3", "bar3"));
@@ -2851,13 +2881,13 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
      // BF is cappped:3 now
      std::unique_ptr<Iterator> iter_tmp(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter_tmp, "foo"), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 4);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 2);
      ASSERT_EQ(CountIter(iter_tmp, "gpk"), 0);
      // both counters are incremented because BF is "not changed" for 1 of the
      // 2 SST files, so filter is checked once and found no match.
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 5);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
    }

    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "fixed:2"}}));
@@ -2874,33 +2904,34 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
      std::unique_ptr<Iterator> iter_tmp(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter_tmp, "foo"), 9);
      // the first and last BF are checked
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 7);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 2);
      ASSERT_EQ(CountIter(iter_tmp, "gpk"), 0);
      // only last BF is checked and not found
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 8);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 2);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
    }

-    // iter_old can only see the first SST, so checked plus 1
+    // iter_old can only see the first SST
    ASSERT_EQ(CountIter(iter_old, "foo"), 4);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 9);
-    // iter was created after the first setoptions call so only full filter
-    // will check the filter
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 1);
+    // same with iter, but different prefix extractor
    ASSERT_EQ(CountIter(iter, "foo"), 2);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 10);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 1);

    {
      // keys in all three SSTs are visible to iterator
      // The range of [foo, foz90000] is compatible with (fixed:1) and (fixed:2)
-      // so +2 for checked counter
      std::unique_ptr<Iterator> iter_all(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter_all, "foo"), 9);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 12);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 2);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 2);
      ASSERT_EQ(CountIter(iter_all, "gpk"), 0);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 13);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 3);
+      // FIXME? isn't seek key out of SST range?
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
    }
    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "capped:3"}}));
    ASSERT_EQ(dbfull()->GetOptions().prefix_extractor->AsString(),
@@ -2910,11 +2941,12 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterMultipleSST) {
      ASSERT_EQ(CountIter(iter_all, "foo"), 6);
      // all three SST are checked because the current options has the same as
      // the remaining SST (capped:3)
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 16);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 3);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 3);
      ASSERT_EQ(CountIter(iter_all, "gpk"), 0);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 17);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 4);
+      // FIXME? isn't seek key out of SST range?
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 1);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
    }
    // TODO(Zhongyi): Maybe also need to add Get calls to test point look up?
  }
@@ -3016,13 +3048,13 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterOptions) {
    {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      ASSERT_EQ(CountIter(iter, "foo"), 12);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 3);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 3);
    }
    std::unique_ptr<Iterator> iter_old(db_->NewIterator(read_options));
    ASSERT_EQ(CountIter(iter_old, "foo"), 12);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 6);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 3);

    ASSERT_OK(dbfull()->SetOptions({{"prefix_extractor", "capped:3"}}));
    ASSERT_EQ(dbfull()->GetOptions().prefix_extractor->AsString(),
@@ -3031,17 +3063,18 @@ TEST_F(DBBloomFilterTest, DynamicBloomFilterOptions) {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      // "fp*" should be skipped
      ASSERT_EQ(CountIter(iter, "foo"), 9);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 6);
-      ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+      EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
    }

    // iterator created before should not be affected and see all keys
    ASSERT_EQ(CountIter(iter_old, "foo"), 12);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 9);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 0);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 3);
    ASSERT_EQ(CountIter(iter_old, "abc"), 0);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED), 12);
-    ASSERT_EQ(TestGetTickerCount(options, BLOOM_FILTER_PREFIX_USEFUL), 3);
+    // FIXME? isn't seek key out of SST range?
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTERED), 3);
+    EXPECT_EQ(PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH), 0);
  }
 }

@@ -3136,19 +3169,19 @@ class FixedSuffix4Transform : public SliceTransform {

 std::pair<uint64_t, uint64_t> GetBloomStat(const Options& options, bool sst) {
  if (sst) {
-    return {
-        options.statistics->getAndResetTickerCount(BLOOM_FILTER_PREFIX_CHECKED),
-        options.statistics->getAndResetTickerCount(BLOOM_FILTER_PREFIX_USEFUL)};
+    return {options.statistics->getAndResetTickerCount(
+                NON_LAST_LEVEL_SEEK_FILTER_MATCH),
+            options.statistics->getAndResetTickerCount(
+                NON_LAST_LEVEL_SEEK_FILTERED)};
  } else {
    auto hit = std::exchange(get_perf_context()->bloom_memtable_hit_count, 0);
    auto miss = std::exchange(get_perf_context()->bloom_memtable_miss_count, 0);
-    return {hit + miss, miss};
+    return {hit, miss};
  }
 }

-std::pair<uint64_t, uint64_t> CheckedAndUseful(uint64_t checked,
-                                               uint64_t useful) {
-  return {checked, useful};
+std::pair<uint64_t, uint64_t> HitAndMiss(uint64_t hits, uint64_t misses) {
+  return {hits, misses};
 }
 }  // anonymous namespace

@@ -3186,27 +3219,27 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter1) {
    if (flushed) {  // TODO: support auto_prefix_mode in memtable?
      read_options.auto_prefix_mode = true;
    }
-    EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+    EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
    {
      Slice ub("999aaaa");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaaa"), 3);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      Slice ub("999abaa");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "abaa"), 1);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      Slice ub("999acaa");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "acaa"), 0);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 1));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 1));
    }
    {
      Slice ub("zzzz");
@@ -3214,7 +3247,7 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter1) {
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "baa"), 3);
      if (flushed) {  // TODO: fix memtable case
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      }
    }
  }
@@ -3260,13 +3293,13 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter2) {
      get_perf_context()->bloom_memtable_hit_count = 0;
      get_perf_context()->bloom_memtable_miss_count = 0;
    }
-    EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+    EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
    {
      Slice ub("aaaa000");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaaa999"), 3);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      // Note: prefix does work as upper bound
@@ -3274,7 +3307,7 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter2) {
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaaa999"), 3);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      // Note: prefix does not work here as seek key
@@ -3282,28 +3315,28 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter2) {
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaaa"), 0);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      Slice ub("aaba000");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaba999"), 1);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
    {
      Slice ub("aaca000");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "aaca999"), 0);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 1));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 1));
    }
    {
      Slice ub("aaaz");
      read_options.iterate_upper_bound = &ub;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "zzz"), 5);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
    }
    {
      // Note: prefix does work here as seek key, but only finds key equal
@@ -3313,7 +3346,7 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter2) {
      read_options.prefix_same_as_start = true;
      std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
      EXPECT_EQ(CountIter(iter, "qqqq"), 1);
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
    }
  }
 }
@@ -3404,13 +3437,13 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter3) {
        get_perf_context()->bloom_memtable_hit_count = 0;
        get_perf_context()->bloom_memtable_miss_count = 0;
      }
-      EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+      EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      {
        Slice ub("aaaa999");
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "aaaa000"), 3);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
      }
      {
        // Note: prefix as seek key is not bloom-optimized
@@ -3420,28 +3453,28 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter3) {
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "aaaa"), 3);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      }
      {
        Slice ub("aaba9");
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "aaba0"), 1);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
      }
      {
        Slice ub("aaca9");
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "aaca0"), 0);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 1));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 1));
      }
      {
        Slice ub("qqqq9");
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "qqqq0"), 1);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(1, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(1, 0));
      }
      {
        // Note: prefix as seek key is not bloom-optimized
@@ -3449,7 +3482,7 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter3) {
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "qqqq"), weird_comparator ? 7 : 2);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      }
      {
        // Note: prefix as seek key is not bloom-optimized
@@ -3457,14 +3490,14 @@ TEST_F(DBBloomFilterTest, WeirdPrefixExtractorWithFilter3) {
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "zzzz"), weird_comparator ? 8 : 1);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      }
      {
        Slice ub("zzzz9");
        read_options.iterate_upper_bound = &ub;
        std::unique_ptr<Iterator> iter(db_->NewIterator(read_options));
        EXPECT_EQ(CountIter(iter, "aab"), weird_comparator ? 6 : 5);
-        EXPECT_EQ(GetBloomStat(options, flushed), CheckedAndUseful(0, 0));
+        EXPECT_EQ(GetBloomStat(options, flushed), HitAndMiss(0, 0));
      }
    }
  }
@@ -0,0 +1,142 @@
+//  Copyright (c) Meta Platforms, Inc. and affiliates.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+
+#include "db/db_test_util.h"
+#include "port/port.h"
+#include "util/random.h"
+
+namespace ROCKSDB_NAMESPACE {
+
+class DBClipTest : public DBTestBase {
+ public:
+  DBClipTest() : DBTestBase("db_clip_test", /*env_do_fsync=*/true) {}
+};
+
+TEST_F(DBClipTest, TestClipRange) {
+  Options options = CurrentOptions();
+  options.write_buffer_size = 10 * 1024 * 1024;
+  options.max_bytes_for_level_multiplier = 2;
+  options.num_levels = 3;
+  options.max_background_compactions = 3;
+  options.disable_auto_compactions = true;
+  options.statistics = CreateDBStatistics();
+
+  DestroyAndReopen(options);
+  int32_t value_size = 10 * 1024;  // 10 KB
+
+  Random rnd(301);
+  std::map<int32_t, std::string> values;
+
+  // file [0 => 100), [100 => 200), ... [900, 1000)
+  for (auto i = 0; i < 10; i++) {
+    for (auto j = 0; j < 100; j++) {
+      auto k = i * 100 + j;
+      values[k] = rnd.RandomString(value_size);
+      ASSERT_OK(Put(Key(k), values[k]));
+    }
+    ASSERT_OK(Flush());
+  }
+  ASSERT_EQ("10", FilesPerLevel(0));
+  auto begin_key = Key(251), end_key = Key(751);
+  ASSERT_OK(
+      db_->ClipColumnFamily(db_->DefaultColumnFamily(), begin_key, end_key));
+
+  for (auto i = 0; i < 251; i++) {
+    ReadOptions ropts;
+    std::string result;
+    auto s = db_->Get(ropts, Key(i), &result);
+    ASSERT_TRUE(s.IsNotFound());
+  }
+  for (auto i = 251; i < 751; i++) {
+    ASSERT_EQ(Get(Key(i)), values[i]);
+  }
+  for (auto i = 751; i < 1000; i++) {
+    ReadOptions ropts;
+    std::string result;
+    auto s = db_->Get(ropts, Key(i), &result);
+    ASSERT_TRUE(s.IsNotFound());
+  }
+
+  std::vector<LiveFileMetaData> all_metadata;
+  db_->GetLiveFilesMetaData(&all_metadata);
+  for (auto& md : all_metadata) {
+    // make sure clip_begin_key <= file_smallestkey <= file_largestkey <=
+    // clip_end_key
+    bool in_range = false;
+
+    if (options.comparator->Compare(begin_key, md.smallestkey) <= 0 &&
+        options.comparator->Compare(end_key, md.largestkey) > 0) {
+      in_range = true;
+    }
+    ASSERT_TRUE(in_range);
+  }
+
+  CompactRangeOptions compact_options;
+  compact_options.change_level = true;
+  compact_options.target_level = 2;
+  ASSERT_OK(db_->CompactRange(compact_options, nullptr, nullptr));
+  ASSERT_EQ("0,0,3", FilesPerLevel(0));
+
+  for (auto i = 0; i < 10; i += 2) {
+    for (auto j = 0; j < 100; j++) {
+      auto k = i * 100 + j;
+      ASSERT_OK(Put(Key(k), values[k]));
+    }
+    ASSERT_OK(Flush());
+  }
+  ASSERT_EQ("5,0,3", FilesPerLevel(0));
+  ASSERT_OK(dbfull()->TEST_CompactRange(0, nullptr, nullptr));
+  ASSERT_EQ("0,5,3", FilesPerLevel(0));
+
+  for (auto i = 1; i < 10; i += 2) {
+    for (auto j = 0; j < 100; j++) {
+      auto k = i * 100 + j;
+      ASSERT_OK(Put(Key(k), values[k]));
+    }
+    ASSERT_OK(Flush());
+  }
+  ASSERT_EQ("5,5,3", FilesPerLevel(0));
+
+  auto begin_key_2 = Key(222), end_key_2 = Key(888);
+
+  ASSERT_OK(db_->ClipColumnFamily(db_->DefaultColumnFamily(), begin_key_2,
+                                  end_key_2));
+
+  for (auto i = 0; i < 222; i++) {
+    ReadOptions ropts;
+    std::string result;
+    auto s = db_->Get(ropts, Key(i), &result);
+    ASSERT_TRUE(s.IsNotFound());
+  }
+  for (auto i = 222; i < 888; i++) {
+    ASSERT_EQ(Get(Key(i)), values[i]);
+  }
+  for (auto i = 888; i < 1000; i++) {
+    ReadOptions ropts;
+    std::string result;
+    auto s = db_->Get(ropts, Key(i), &result);
+    ASSERT_TRUE(s.IsNotFound());
+  }
+
+  std::vector<LiveFileMetaData> all_metadata_2;
+  db_->GetLiveFilesMetaData(&all_metadata_2);
+  for (auto& md : all_metadata_2) {
+    // make sure clip_begin_key <= file_smallestkey <= file_largestkey <=
+    // clip_end_key
+    bool in_range = false;
+    if (begin_key_2.compare(md.smallestkey) <= 0 &&
+        end_key_2.compare(md.largestkey) > 0) {
+      in_range = true;
+    }
+    ASSERT_TRUE(in_range);
+  }
+}
+}  // namespace ROCKSDB_NAMESPACE
+
+int main(int argc, char** argv) {
+  ROCKSDB_NAMESPACE::port::InstallStackTraceHandler();
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}
@@ -1105,7 +1105,7 @@ TEST_F(DBCompactionTest, CompactionSstPartitionerNonTrivial) {
  ASSERT_OK(Put("bbbb1", "B"));
  ASSERT_OK(Flush());
  ASSERT_OK(dbfull()->TEST_WaitForFlushMemTable());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  std::vector<LiveFileMetaData> files;
  dbfull()->GetLiveFilesMetaData(&files);
@@ -5075,7 +5075,11 @@ TEST_F(DBCompactionTest, CompactRangeShutdownWhileDelayed) {
    manual_compaction_thread.join();
    TEST_SYNC_POINT(
        "DBCompactionTest::CompactRangeShutdownWhileDelayed:PostManual");
-    ASSERT_OK(dbfull()->TEST_WaitForCompact());
+    if (i == 0) {
+      ASSERT_OK(dbfull()->TEST_WaitForCompact());
+    } else {
+      ASSERT_NOK(dbfull()->TEST_WaitForCompact());
+    }
    ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
  }
 }
@@ -6740,10 +6744,8 @@ class DBCompactionTestL0FilesMisorderCorruption : public DBCompactionTest {
      if (compaction_path_to_test == "FindIntraL0Compaction" ||
          compaction_path_to_test == "CompactRange") {
        fifo_options.allow_compaction = true;
-        fifo_options.age_for_warm = 0;
      } else if (compaction_path_to_test == "CompactFile") {
        fifo_options.allow_compaction = false;
-        fifo_options.age_for_warm = 0;
      }
      options_.compaction_options_fifo = fifo_options;
    }
@@ -8593,7 +8595,7 @@ TEST_F(DBCompactionTest, CompactionWithChecksumHandoffManifest2) {
  Destroy(options);
 }

-TEST_F(DBCompactionTest, FIFOWarm) {
+TEST_F(DBCompactionTest, FIFOChangeTemperature) {
  Options options = CurrentOptions();
  options.compaction_style = kCompactionStyleFIFO;
  options.num_levels = 1;
@@ -8601,18 +8603,18 @@ TEST_F(DBCompactionTest, FIFOWarm) {
  options.level0_file_num_compaction_trigger = 2;
  options.create_if_missing = true;
  CompactionOptionsFIFO fifo_options;
-  fifo_options.age_for_warm = 1000;
+  fifo_options.file_temperature_age_thresholds = {{Temperature::kCold, 1000}};
  fifo_options.max_table_files_size = 100000000;
  options.compaction_options_fifo = fifo_options;
  env_->SetMockSleep();
  Reopen(options);

-  int total_warm = 0;
+  int total_cold = 0;
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
      "NewWritableFile::FileOptions.temperature", [&](void* arg) {
        Temperature temperature = *(static_cast<Temperature*>(arg));
-        if (temperature == Temperature::kWarm) {
-          total_warm++;
+        if (temperature == Temperature::kCold) {
+          total_cold++;
        }
      });
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
@@ -8649,9 +8651,9 @@ TEST_F(DBCompactionTest, FIFOWarm) {
  ASSERT_EQ(4, metadata.file_count);
  ASSERT_EQ(Temperature::kUnknown, metadata.levels[0].files[0].temperature);
  ASSERT_EQ(Temperature::kUnknown, metadata.levels[0].files[1].temperature);
-  ASSERT_EQ(Temperature::kWarm, metadata.levels[0].files[2].temperature);
-  ASSERT_EQ(Temperature::kWarm, metadata.levels[0].files[3].temperature);
-  ASSERT_EQ(2, total_warm);
+  ASSERT_EQ(Temperature::kCold, metadata.levels[0].files[2].temperature);
+  ASSERT_EQ(Temperature::kCold, metadata.levels[0].files[3].temperature);
+  ASSERT_EQ(2, total_cold);

  Destroy(options);
 }
@@ -8716,7 +8718,7 @@ TEST_F(DBCompactionTest, DisableMultiManualCompaction) {

  sleeping_task_low.WakeUp();
  sleeping_task_low.WaitUntilDone();
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
 }

 TEST_F(DBCompactionTest, DisableJustStartedManualCompaction) {
@@ -8848,7 +8850,7 @@ TEST_F(DBCompactionTest, DisableManualCompactionThreadQueueFull) {

  sleeping_task_low.WakeUp();
  sleeping_task_low.WaitUntilDone();
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,1", FilesPerLevel(0));
 }

@@ -128,6 +128,13 @@ class CompactedDBImpl : public DBImpl {
    return Status::NotSupported("Not supported in compacted db mode.");
  }

+  using DB::ClipColumnFamily;
+  virtual Status ClipColumnFamily(ColumnFamilyHandle* /*column_family*/,
+                                  const Slice& /*begin*/,
+                                  const Slice& /*end*/) override {
+    return Status::NotSupported("Not supported in compacted db mode.");
+  }
+
  // FIXME: some missing overrides for more "write" functions
  // Share with DBImplReadOnly?

@@ -258,6 +258,9 @@ DBImpl::DBImpl(const DBOptions& options, const std::string& dbname,
  co.capacity = table_cache_size;
  co.num_shard_bits = immutable_db_options_.table_cache_numshardbits;
  co.metadata_charge_policy = kDontChargeCacheMetadata;
+  // TODO: Consider a non-fixed seed once test fallout (prefetch_test) is
+  // dealt with
+  co.hash_seed = 0;
  table_cache_ = NewLRUCache(co);
  SetDbSessionId();
  assert(!db_session_id_.empty());
@@ -2475,7 +2478,6 @@ std::vector<Status> DBImpl::MultiGet(

  // Post processing (decrement reference counts and record statistics)
  PERF_TIMER_GUARD(get_post_process_time);
-  autovector<SuperVersion*> superversions_to_delete;

  for (auto mgd_iter : multiget_cf_data) {
    auto mgd = mgd_iter.second;
@@ -4520,7 +4522,6 @@ void DBImpl::GetAllColumnFamilyMetaData(
  }
 }

-
 Status DBImpl::CheckConsistency() {
  mutex_.AssertHeld();
  std::vector<LiveFileMetaData> metadata;
@@ -4911,6 +4912,15 @@ Status DBImpl::WriteOptionsFile(bool need_mutex_lock,
  if (s.ok()) {
    s = RenameTempFileToOptionsFile(file_name);
  }
+
+  if (!s.ok() && GetEnv()->FileExists(file_name).ok()) {
+    if (!GetEnv()->DeleteFile(file_name).ok()) {
+      ROCKS_LOG_WARN(immutable_db_options_.info_log,
+                     "Unable to delete temp options file %s",
+                     file_name.c_str());
+    }
+  }
+
  // restore lock
  if (!need_mutex_lock) {
    mutex_.Lock();
@@ -5694,6 +5704,81 @@ Status DBImpl::CreateColumnFamilyWithImport(
  return status;
 }

+Status DBImpl::ClipColumnFamily(ColumnFamilyHandle* column_family,
+                                const Slice& begin_key, const Slice& end_key) {
+  assert(column_family);
+  Status status;
+  // Flush memtable
+  FlushOptions flush_opts;
+  flush_opts.allow_write_stall = true;
+  auto* cfd =
+      static_cast_with_check<ColumnFamilyHandleImpl>(column_family)->cfd();
+  if (immutable_db_options_.atomic_flush) {
+    status = AtomicFlushMemTables(flush_opts, FlushReason::kDeleteFiles,
+                                  {} /* provided_candidate_cfds */,
+                                  false /* entered_write_thread */);
+  } else {
+    status = FlushMemTable(cfd, flush_opts, FlushReason::kDeleteFiles,
+                           false /* entered_write_thread */);
+  }
+
+  if (status.ok()) {
+    // DeleteFilesInRanges non-overlap files except L0
+    std::vector<RangePtr> ranges;
+    ranges.push_back(RangePtr(nullptr, &begin_key));
+    ranges.push_back(RangePtr(&end_key, nullptr));
+    status = DeleteFilesInRanges(column_family, ranges.data(), ranges.size());
+  }
+
+  // DeleteRange the remaining overlapping keys
+  bool empty_after_delete = false;
+  if (status.ok()) {
+    Slice smallest_user_key, largest_user_key;
+    {
+      // Lock db mutex
+      InstrumentedMutexLock l(&mutex_);
+      cfd->current()->GetSstFilesBoundaryKeys(&smallest_user_key,
+                                              &largest_user_key);
+    }
+    // all the files has been deleted after DeleteFilesInRanges;
+    if (smallest_user_key.empty() && largest_user_key.empty()) {
+      empty_after_delete = true;
+    } else {
+      const Comparator* const ucmp = column_family->GetComparator();
+      WriteOptions wo;
+      // Delete [smallest_user_key, clip_begin_key)
+      if (ucmp->Compare(smallest_user_key, begin_key) < 0) {
+        status = DeleteRange(wo, column_family, smallest_user_key, begin_key);
+      }
+
+      if (status.ok()) {
+        // Delete [clip_end_key, largest_use_key]
+        if (ucmp->Compare(end_key, largest_user_key) < 0) {
+          status = DeleteRange(wo, column_family, end_key, largest_user_key);
+          if (status.ok()) {
+            status = Delete(wo, column_family, largest_user_key);
+          }
+        }
+      }
+    }
+  }
+
+  if (status.ok() && !empty_after_delete) {
+    // CompactRange delete all the tombstones
+    CompactRangeOptions compact_options;
+    compact_options.exclusive_manual_compaction = true;
+    compact_options.bottommost_level_compaction =
+        BottommostLevelCompaction::kForceOptimized;
+    // We could just compact the ranges [null, clip_begin_key] and
+    // [clip_end_key, null]. But due to how manual compaction calculates the
+    // last level to compact to and that range tombstones are not dropped
+    // during non-bottommost compactions, calling CompactRange() on these two
+    // ranges may not clear all range tombstones.
+    status = CompactRange(compact_options, nullptr, nullptr);
+  }
+  return status;
+}
+
 Status DBImpl::VerifyFileChecksums(const ReadOptions& read_options) {
  return VerifyChecksumInternal(read_options, /*use_file_checksum=*/true);
 }
@@ -534,6 +534,11 @@ class DBImpl : public DB {
      const ExportImportFilesMetaData& metadata,
      ColumnFamilyHandle** handle) override;

+  using DB::ClipColumnFamily;
+  virtual Status ClipColumnFamily(ColumnFamilyHandle* column_family,
+                                  const Slice& begin_key,
+                                  const Slice& end_key) override;
+
  using DB::VerifyFileChecksums;
  Status VerifyFileChecksums(const ReadOptions& read_options) override;

@@ -1053,10 +1058,16 @@ class DBImpl : public DB {

  VersionSet* GetVersionSet() const { return versions_.get(); }

-  // Wait for any compaction
-  // We add a bool parameter to wait for unscheduledCompactions_ == 0, but this
-  // is only for the special test of CancelledCompactions
-  Status WaitForCompact(bool waitUnscheduled = false);
+  // Wait for all flush and compactions jobs to finish. Jobs to wait include the
+  // unscheduled (queued, but not scheduled yet). If the db is shutting down,
+  // Status::ShutdownInProgress will be returned. If PauseBackgroundWork() was
+  // called prior to this, this may potentially wait for unscheduled jobs
+  // indefinitely. abort_on_pause can be set to true to abort, and
+  // Status::Aborted will be returned immediately. This may also never return if
+  // there's sufficient ongoing writes that keeps flush and compaction going
+  // without stopping. The user would have to cease all the writes to DB to make
+  // this eventually return in a stable state.
+  Status WaitForCompact(bool abort_on_pause = false);

 #ifndef NDEBUG
  // Compact any files in the named level that overlap [*begin, *end]
@@ -1096,9 +1107,7 @@ class DBImpl : public DB {
  Status TEST_WaitForFlushMemTable(ColumnFamilyHandle* column_family = nullptr);

  // Wait for any compaction
-  // We add a bool parameter to wait for unscheduledCompactions_ == 0, but this
-  // is only for the special test of CancelledCompactions
-  Status TEST_WaitForCompact(bool waitUnscheduled = false);
+  Status TEST_WaitForCompact(bool abort_on_pause = false);

  // Wait for any background purge
  Status TEST_WaitForPurge();
@@ -2310,7 +2319,7 @@ class DBImpl : public DB {
  // logfile_number_. With two_write_queues it also protects alive_log_files_,
  // and log_empty_. Refer to the definition of each variable below for more
  // details.
-  // Note: to avoid dealock, if needed to acquire both log_write_mutex_ and
+  // Note: to avoid deadlock, if needed to acquire both log_write_mutex_ and
  // mutex_, the order should be first mutex_ and then log_write_mutex_.
  InstrumentedMutex log_write_mutex_;

@@ -1758,7 +1758,7 @@ Status DBImpl::ReFitLevel(ColumnFamilyData* cfd, int level, int target_level) {
          f->marked_for_compaction, f->temperature, f->oldest_blob_file_number,
          f->oldest_ancester_time, f->file_creation_time, f->epoch_number,
          f->file_checksum, f->file_checksum_func_name, f->unique_id,
-          f->compensated_range_deletion_size);
+          f->compensated_range_deletion_size, f->tail_size);
    }
    ROCKS_LOG_DEBUG(immutable_db_options_.info_log,
                    "[%s] Apply version edit:\n%s", cfd->GetName().c_str(),
@@ -3457,7 +3457,7 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
            f->oldest_blob_file_number, f->oldest_ancester_time,
            f->file_creation_time, f->epoch_number, f->file_checksum,
            f->file_checksum_func_name, f->unique_id,
-            f->compensated_range_deletion_size);
+            f->compensated_range_deletion_size, f->tail_size);

        ROCKS_LOG_BUFFER(
            log_buffer,
@@ -3569,7 +3569,7 @@ Status DBImpl::BackgroundCompaction(bool* made_progress,
    mutex_.Unlock();
    TEST_SYNC_POINT_CALLBACK(
        "DBImpl::BackgroundCompaction:NonTrivial:BeforeRun", nullptr);
-    // Should handle erorr?
+    // Should handle error?
    compaction_job.Run().PermitUncheckedError();
    TEST_SYNC_POINT("DBImpl::BackgroundCompaction:NonTrivial:AfterRun");
    mutex_.Lock();
@@ -3957,16 +3957,24 @@ void DBImpl::GetSnapshotContext(
  *snapshot_seqs = snapshots_.GetAll(earliest_write_conflict_snapshot);
 }

-Status DBImpl::WaitForCompact(bool wait_unscheduled) {
-  // Wait until the compaction completes
+Status DBImpl::WaitForCompact(bool abort_on_pause) {
  InstrumentedMutexLock l(&mutex_);
-  while ((bg_bottom_compaction_scheduled_ || bg_compaction_scheduled_ ||
-          bg_flush_scheduled_ ||
-          (wait_unscheduled && unscheduled_compactions_)) &&
-         (error_handler_.GetBGError().ok())) {
-    bg_cv_.Wait();
+  for (;;) {
+    if (shutting_down_.load(std::memory_order_acquire)) {
+      return Status::ShutdownInProgress();
+    }
+    if (bg_work_paused_ && abort_on_pause) {
+      return Status::Aborted();
+    }
+    if ((bg_bottom_compaction_scheduled_ || bg_compaction_scheduled_ ||
+         bg_flush_scheduled_ || unscheduled_compactions_ ||
+         unscheduled_flushes_) &&
+        (error_handler_.GetBGError().ok())) {
+      bg_cv_.Wait();
+    } else {
+      return error_handler_.GetBGError();
+    }
  }
-  return error_handler_.GetBGError();
 }

 }  // namespace ROCKSDB_NAMESPACE
@@ -178,9 +178,9 @@ Status DBImpl::TEST_WaitForFlushMemTable(ColumnFamilyHandle* column_family) {
  return WaitForFlushMemTable(cfd, nullptr, false);
 }

-Status DBImpl::TEST_WaitForCompact(bool wait_unscheduled) {
+Status DBImpl::TEST_WaitForCompact(bool abort_on_pause) {
  // Wait until the compaction completes
-  return WaitForCompact(wait_unscheduled);
+  return WaitForCompact(abort_on_pause);
 }

 Status DBImpl::TEST_WaitForPurge() {
@@ -138,7 +138,7 @@ Status DBImpl::PromoteL0(ColumnFamilyHandle* column_family, int target_level) {
                   f->oldest_blob_file_number, f->oldest_ancester_time,
                   f->file_creation_time, f->epoch_number, f->file_checksum,
                   f->file_checksum_func_name, f->unique_id,
-                   f->compensated_range_deletion_size);
+                   f->compensated_range_deletion_size, f->tail_size);
    }

    status = versions_->LogAndApply(cfd, *cfd->GetLatestMutableCFOptions(),
@@ -286,6 +286,7 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
    return;
  }

+  bool mutex_unlocked = false;
  if (!alive_log_files_.empty() && !logs_.empty()) {
    uint64_t min_log_number = job_context->log_number;
    size_t num_alive_log_files = alive_log_files_.size();
@@ -315,6 +316,7 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,
    }
    log_write_mutex_.Unlock();
    mutex_.Unlock();
+    mutex_unlocked = true;
    TEST_SYNC_POINT_CALLBACK("FindObsoleteFiles::PostMutexUnlock", nullptr);
    log_write_mutex_.Lock();
    while (!logs_.empty() && logs_.front().number < min_log_number) {
@@ -337,7 +339,9 @@ void DBImpl::FindObsoleteFiles(JobContext* job_context, bool force,

  logs_to_free_.clear();
  log_write_mutex_.Unlock();
-  mutex_.Lock();
+  if (mutex_unlocked) {
+    mutex_.Lock();
+  }
  job_context->log_recycle_files.assign(log_recycle_files_.begin(),
                                        log_recycle_files_.end());
 }
@@ -24,7 +24,7 @@
 #include "rocksdb/table.h"
 #include "rocksdb/wal_filter.h"
 #include "test_util/sync_point.h"
-#include "util/rate_limiter.h"
+#include "util/rate_limiter_impl.h"

 namespace ROCKSDB_NAMESPACE {
 Options SanitizeOptions(const std::string& dbname, const Options& src,
@@ -616,7 +616,8 @@ Status DBImpl::Recover(
                           f->oldest_blob_file_number, f->oldest_ancester_time,
                           f->file_creation_time, f->epoch_number,
                           f->file_checksum, f->file_checksum_func_name,
-                           f->unique_id, f->compensated_range_deletion_size);
+                           f->unique_id, f->compensated_range_deletion_size,
+                           f->tail_size);
              ROCKS_LOG_WARN(immutable_db_options_.info_log,
                             "[%s] Moving #%" PRIu64
                             " from from_level-%d to from_level-%d %" PRIu64
@@ -1453,7 +1454,7 @@ Status DBImpl::RecoverLogFiles(const std::vector<uint64_t>& wal_numbers,
        recovery_ctx->UpdateVersionEdits(cfd, iter->second);
      }

-      if (flushed) {
+      if (flushed || !data_seen) {
        VersionEdit wal_deletion;
        if (immutable_db_options_.track_and_verify_wals_in_manifest) {
          wal_deletion.DeleteWalsBefore(max_wal_number + 1);
@@ -1673,14 +1674,14 @@ Status DBImpl::WriteLevel0TableForRecovery(int job_id, ColumnFamilyData* cfd,
  constexpr int level = 0;

  if (s.ok() && has_output) {
-    edit->AddFile(level, meta.fd.GetNumber(), meta.fd.GetPathId(),
-                  meta.fd.GetFileSize(), meta.smallest, meta.largest,
-                  meta.fd.smallest_seqno, meta.fd.largest_seqno,
-                  meta.marked_for_compaction, meta.temperature,
-                  meta.oldest_blob_file_number, meta.oldest_ancester_time,
-                  meta.file_creation_time, meta.epoch_number,
-                  meta.file_checksum, meta.file_checksum_func_name,
-                  meta.unique_id, meta.compensated_range_deletion_size);
+    edit->AddFile(
+        level, meta.fd.GetNumber(), meta.fd.GetPathId(), meta.fd.GetFileSize(),
+        meta.smallest, meta.largest, meta.fd.smallest_seqno,
+        meta.fd.largest_seqno, meta.marked_for_compaction, meta.temperature,
+        meta.oldest_blob_file_number, meta.oldest_ancester_time,
+        meta.file_creation_time, meta.epoch_number, meta.file_checksum,
+        meta.file_checksum_func_name, meta.unique_id,
+        meta.compensated_range_deletion_size, meta.tail_size);

    for (const auto& blob : blob_file_additions) {
      edit->AddBlobFile(blob);
@@ -142,6 +142,13 @@ class DBImplReadOnly : public DBImpl {
    return Status::NotSupported("Not supported operation in read only mode.");
  }

+  using DB::ClipColumnFamily;
+  virtual Status ClipColumnFamily(ColumnFamilyHandle* /*column_family*/,
+                                  const Slice& /*begin*/,
+                                  const Slice& /*end*/) override {
+    return Status::NotSupported("Not supported operation in read only mode.");
+  }
+
  // FIXME: some missing overrides for more "write" functions

 protected:
@@ -964,9 +964,6 @@ bool DBIter::FindValueForCurrentKey() {
    assert(last_key_entry_type == ikey_.type);
  }

-  Status s;
-  s.PermitUncheckedError();
-
  switch (last_key_entry_type) {
    case kTypeDeletion:
    case kTypeDeletionWithTimestamp:
@@ -1043,11 +1040,6 @@ bool DBIter::FindValueForCurrentKey() {
          std::to_string(static_cast<unsigned int>(last_key_entry_type)));
      return false;
  }
-  if (!s.ok()) {
-    valid_ = false;
-    status_ = s;
-    return false;
-  }
  valid_ = true;
  return true;
 }
@@ -16,7 +16,7 @@
 #include "port/stack_trace.h"
 #include "rocksdb/iostats_context.h"
 #include "rocksdb/perf_context.h"
-#include "table/block_based/flush_block_policy.h"
+#include "table/block_based/flush_block_policy_impl.h"
 #include "util/random.h"
 #include "utilities/merge_operators/string_append/stringappend2.h"

@@ -22,6 +22,7 @@
 #include "test_util/sync_point.h"
 #include "test_util/testutil.h"
 #include "util/random.h"
+#include "utilities/fault_injection_fs.h"

 namespace ROCKSDB_NAMESPACE {

@@ -910,6 +911,7 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
  options.compression = kNoCompression;
  options.create_if_missing = true;
  options.compaction_options_fifo.allow_compaction = false;
+  options.num_levels = 1;
  env_->SetMockSleep();
  options.env = env_;

@@ -1009,6 +1011,24 @@ TEST_F(DBOptionsTest, SetFIFOCompactionOptions) {
  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_GE(NumTableFilesAtLevel(0), 1);
  ASSERT_LE(NumTableFilesAtLevel(0), 5);
+
+  // Test dynamically setting `file_temperature_age_thresholds`
+  ASSERT_TRUE(
+      dbfull()
+          ->GetOptions()
+          .compaction_options_fifo.file_temperature_age_thresholds.empty());
+  ASSERT_OK(dbfull()->SetOptions({{"compaction_options_fifo",
+                                   "{file_temperature_age_thresholds={{age=10;"
+                                   "temperature=kWarm}:{age=30000;"
+                                   "temperature=kCold}}}"}}));
+  auto opts = dbfull()->GetOptions();
+  const auto& fifo_temp_opt =
+      opts.compaction_options_fifo.file_temperature_age_thresholds;
+  ASSERT_EQ(fifo_temp_opt.size(), 2);
+  ASSERT_EQ(fifo_temp_opt[0].temperature, Temperature::kWarm);
+  ASSERT_EQ(fifo_temp_opt[0].age, 10);
+  ASSERT_EQ(fifo_temp_opt[1].temperature, Temperature::kCold);
+  ASSERT_EQ(fifo_temp_opt[1].age, 30000);
 }

 TEST_F(DBOptionsTest, CompactionReadaheadSizeChange) {
@@ -1043,6 +1063,7 @@ TEST_F(DBOptionsTest, FIFOTtlBackwardCompatible) {
  options.write_buffer_size = 10 << 10;  // 10KB
  options.create_if_missing = true;
  options.env = CurrentOptions().env;
+  options.num_levels = 1;

  ASSERT_OK(TryReopen(options));

@@ -1063,12 +1084,19 @@ TEST_F(DBOptionsTest, FIFOTtlBackwardCompatible) {
  // ttl under compaction_options_fifo.
  ASSERT_OK(dbfull()->SetOptions(
      {{"compaction_options_fifo",
-        "{allow_compaction=true;max_table_files_size=1024;ttl=731;}"},
+        "{allow_compaction=true;max_table_files_size=1024;ttl=731;file_"
+        "temperature_age_thresholds={temperature=kCold;age=12345}}"},
       {"ttl", "60"}}));
  ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.allow_compaction,
            true);
  ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.max_table_files_size,
            1024);
+  auto opts = dbfull()->GetOptions();
+  const auto& file_temp_age =
+      opts.compaction_options_fifo.file_temperature_age_thresholds;
+  ASSERT_EQ(file_temp_age.size(), 1);
+  ASSERT_EQ(file_temp_age[0].temperature, Temperature::kCold);
+  ASSERT_EQ(file_temp_age[0].age, 12345);
  ASSERT_EQ(dbfull()->GetOptions().ttl, 60);

  // Put ttl as the first option inside compaction_options_fifo. That works as
@@ -1081,6 +1109,9 @@ TEST_F(DBOptionsTest, FIFOTtlBackwardCompatible) {
            true);
  ASSERT_EQ(dbfull()->GetOptions().compaction_options_fifo.max_table_files_size,
            1024);
+  ASSERT_EQ(file_temp_age.size(), 1);
+  ASSERT_EQ(file_temp_age[0].temperature, Temperature::kCold);
+  ASSERT_EQ(file_temp_age[0].age, 12345);
  ASSERT_EQ(dbfull()->GetOptions().ttl, 191);
 }

@@ -1206,6 +1237,90 @@ TEST_F(DBOptionsTest, BottommostCompressionOptsWithFallbackType) {
  ASSERT_EQ(kBottommostCompressionLevel, compression_opt_used.level);
 }

+TEST_F(DBOptionsTest, FIFOTemperatureAgeThresholdValidation) {
+  Options options = CurrentOptions();
+  Destroy(options);
+
+  options.num_levels = 1;
+  options.compaction_style = kCompactionStyleFIFO;
+  options.max_open_files = -1;
+  // elements are not sorted
+  // During DB open
+  options.compaction_options_fifo.file_temperature_age_thresholds.push_back(
+      {Temperature::kCold, 1000});
+  options.compaction_options_fifo.file_temperature_age_thresholds.push_back(
+      {Temperature::kWarm, 500});
+  Status s = TryReopen(options);
+  ASSERT_TRUE(s.IsNotSupported());
+  ASSERT_TRUE(std::strstr(
+      s.getState(),
+      "Option file_temperature_age_thresholds requires elements to be sorted "
+      "in increasing order with respect to `age` field."));
+  // Dynamically set option
+  options.compaction_options_fifo.file_temperature_age_thresholds.pop_back();
+  ASSERT_OK(TryReopen(options));
+  s = db_->SetOptions({{"compaction_options_fifo",
+                        "{file_temperature_age_thresholds={{temperature=kCold;"
+                        "age=1000000}:{temperature=kWarm;age=1}}}"}});
+  ASSERT_TRUE(s.IsNotSupported());
+  ASSERT_TRUE(std::strstr(
+      s.getState(),
+      "Option file_temperature_age_thresholds requires elements to be sorted "
+      "in increasing order with respect to `age` field."));
+
+  // not single level
+  // During DB open
+  options.num_levels = 2;
+  s = TryReopen(options);
+  ASSERT_TRUE(s.IsNotSupported());
+  ASSERT_TRUE(std::strstr(s.getState(),
+                          "Option file_temperature_age_thresholds is only "
+                          "supported when num_levels = 1."));
+  // Dynamically set option
+  options.compaction_options_fifo.file_temperature_age_thresholds.clear();
+  DestroyAndReopen(options);
+  s = db_->SetOptions(
+      {{"compaction_options_fifo",
+        "{file_temperature_age_thresholds={temperature=kCold;age=1000}}"}});
+  ASSERT_TRUE(s.IsNotSupported());
+  ASSERT_TRUE(std::strstr(s.getState(),
+                          "Option file_temperature_age_thresholds is only "
+                          "supported when num_levels = 1."));
+}
+
+TEST_F(DBOptionsTest, TempOptionsFailTest) {
+  std::shared_ptr<FaultInjectionTestFS> fs;
+  std::unique_ptr<Env> env;
+
+  fs.reset(new FaultInjectionTestFS(env_->GetFileSystem()));
+  env = NewCompositeEnv(fs);
+  Options options = CurrentOptions();
+  options.env = env.get();
+
+  SyncPoint::GetInstance()->SetCallBack(
+      "PersistRocksDBOptions:create",
+      [&](void* /*arg*/) { fs->SetFilesystemActive(false); });
+  SyncPoint::GetInstance()->SetCallBack(
+      "PersistRocksDBOptions:written",
+      [&](void* /*arg*/) { fs->SetFilesystemActive(true); });
+
+  SyncPoint::GetInstance()->EnableProcessing();
+  TryReopen(options);
+  SyncPoint::GetInstance()->DisableProcessing();
+
+  std::vector<std::string> filenames;
+  ASSERT_OK(env_->GetChildren(dbname_, &filenames));
+  uint64_t number;
+  FileType type;
+  bool found_temp_file = false;
+  for (size_t i = 0; i < filenames.size(); i++) {
+    if (ParseFileName(filenames[i], &number, &type) && type == kTempFile) {
+      found_temp_file = true;
+    }
+  }
+  ASSERT_FALSE(found_temp_file);
+}
+
 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
@@ -743,7 +743,7 @@ TEST_P(DBSSTTestRateLimit, RateLimitedDelete) {

  // Compaction will move the 4 files in L0 to trash and create 1 L1 file
  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,1", FilesPerLevel(0));

  uint64_t delete_start_time = env_->NowMicros();
@@ -815,7 +815,7 @@ TEST_F(DBSSTTest, RateLimitedWALDelete) {
  CompactRangeOptions cro;
  cro.bottommost_level_compaction = BottommostLevelCompaction::kForceOptimized;
  ASSERT_OK(db_->CompactRange(cro, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,1", FilesPerLevel(0));

  sfm->WaitForEmptyTrash();
@@ -1224,7 +1224,7 @@ TEST_F(DBSSTTest, CancellingCompactionsWorks) {
    ASSERT_OK(Put(Key(i), rnd.RandomString(50)));
  }
  ASSERT_OK(Flush());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  // Because we set a callback in CancelledCompaction, we actually
  // let the compaction run
@@ -1281,7 +1281,7 @@ TEST_F(DBSSTTest, CancellingManualCompactionsWorks) {
                  .IsCompactionTooLarge());

  // Wait for manual compaction to get scheduled and finish
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_EQ(sfm->GetCompactionsReservedSize(), 0);
  // Make sure the stat is bumped
@@ -1297,7 +1297,7 @@ TEST_F(DBSSTTest, CancellingManualCompactionsWorks) {
          .IsCompactionTooLarge());

  // Wait for manual compaction to get scheduled and finish
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_EQ(dbfull()->immutable_db_options().statistics.get()->getTickerCount(
                COMPACTION_CANCELLED),
@@ -1314,7 +1314,7 @@ TEST_F(DBSSTTest, CancellingManualCompactionsWorks) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
  ASSERT_OK(dbfull()->CompactFiles(ROCKSDB_NAMESPACE::CompactionOptions(),
                                   l0_files, 0));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_EQ(sfm->GetCompactionsReservedSize(), 0);
  ASSERT_GT(completed_compactions, 0);
@@ -242,6 +242,46 @@ TEST_F(DBStatisticsTest, VerifyChecksumReadStat) {
  }
 }

+TEST_F(DBStatisticsTest, BlockChecksumStats) {
+  Options options = CurrentOptions();
+  options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
+  Reopen(options);
+
+  // Scenario 0: only WAL data. Not verified so require ticker to be zero.
+  ASSERT_OK(Put("foo", "value"));
+  ASSERT_OK(db_->VerifyChecksum());
+  ASSERT_EQ(0,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_COMPUTE_COUNT));
+  ASSERT_EQ(0,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_MISMATCH_COUNT));
+
+  // Scenario 1: Flushed table verified in `VerifyChecksum()`. This opens a
+  // `TableReader` to verify each of the four blocks (meta-index, table
+  // properties, index, and data block).
+  ASSERT_OK(Flush());
+  ASSERT_OK(options.statistics->Reset());
+  ASSERT_OK(db_->VerifyChecksum());
+  ASSERT_EQ(4,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_COMPUTE_COUNT));
+  ASSERT_EQ(0,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_MISMATCH_COUNT));
+
+  // Scenario 2: Corrupted table verified in `VerifyChecksum()`. The corruption
+  // is in the fourth and final verified block, i.e., the data block.
+  std::unordered_map<std::string, uint64_t> table_files;
+  ASSERT_OK(GetAllDataFiles(kTableFile, &table_files));
+  ASSERT_EQ(1, table_files.size());
+  std::string table_name = table_files.begin()->first;
+  // Assumes the data block starts at offset zero.
+  ASSERT_OK(test::CorruptFile(options.env, table_name, 0 /* offset */,
+                              3 /* bytes_to_corrupt */));
+  ASSERT_OK(options.statistics->Reset());
+  ASSERT_NOK(db_->VerifyChecksum());
+  ASSERT_EQ(4,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_COMPUTE_COUNT));
+  ASSERT_EQ(1,
+            options.statistics->getTickerCount(BLOCK_CHECKSUM_MISMATCH_COUNT));
+}

 }  // namespace ROCKSDB_NAMESPACE

@@ -66,7 +66,7 @@
 #include "util/compression.h"
 #include "util/mutexlock.h"
 #include "util/random.h"
-#include "util/rate_limiter.h"
+#include "util/rate_limiter_impl.h"
 #include "util/string_util.h"
 #include "utilities/merge_operators.h"

@@ -583,7 +583,7 @@ TEST_F(DBTest, LevelReopenWithFIFO) {
        TryReopenWithColumnFamilies({"default", "pikachu"}, fifo_options));
    // For FIFO to pick a compaction
    ASSERT_OK(dbfull()->TEST_CompactRange(0, nullptr, nullptr, handles_[1]));
-    ASSERT_OK(dbfull()->TEST_WaitForCompact(false));
+    ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
    for (int g = 0; g < kKeyCount; ++g) {
      std::string get_key = std::string(1, char('a' + g));
      int status_index = i / kKeyCount;
@@ -3107,6 +3107,13 @@ class ModelDB : public DB {
    return Status::NotSupported("Not implemented.");
  }

+  using DB::ClipColumnFamily;
+  virtual Status ClipColumnFamily(ColumnFamilyHandle* /*column_family*/,
+                                  const Slice& /*begin*/,
+                                  const Slice& /*end*/) override {
+    return Status::NotSupported("Not implemented.");
+  }
+
  using DB::GetPropertiesOfAllTables;
  Status GetPropertiesOfAllTables(
      ColumnFamilyHandle* /*column_family*/,
@@ -3460,6 +3467,7 @@ static bool CompareIterators(int step, DB* model, DB* db,
      ok = false;
    }
  }
+  (void)count;
  delete miter;
  delete dbiter;
  return ok;
@@ -4936,7 +4944,7 @@ TEST_P(DBTestWithParam, PreShutdownMultipleCompaction) {
  ASSERT_GE(operation_count[ThreadStatus::OP_COMPACTION], 1);
  CancelAllBackgroundWork(db_);
  TEST_SYNC_POINT("DBTest::PreShutdownMultipleCompaction:VerifyPreshutdown");
-  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+  ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
  // Record the number of compactions at a time.
  for (int i = 0; i < ThreadStatus::NUM_OP_TYPES; ++i) {
    operation_count[i] = 0;
@@ -5023,7 +5031,7 @@ TEST_P(DBTestWithParam, PreShutdownCompactionMiddle) {
  CancelAllBackgroundWork(db_);
  TEST_SYNC_POINT("DBTest::PreShutdownCompactionMiddle:Preshutdown");
  TEST_SYNC_POINT("DBTest::PreShutdownCompactionMiddle:VerifyPreshutdown");
-  ASSERT_OK(dbfull()->TEST_WaitForCompact());
+  ASSERT_OK(dbfull()->TEST_WaitForBackgroundWork());
  // Record the number of compactions at a time.
  for (int i = 0; i < ThreadStatus::NUM_OP_TYPES; ++i) {
    operation_count[i] = 0;
@@ -6886,7 +6894,20 @@ TEST_F(DBTest, CreateColumnFamilyShouldFailOnIncompatibleOptions) {
 TEST_F(DBTest, RowCache) {
  Options options = CurrentOptions();
  options.statistics = ROCKSDB_NAMESPACE::CreateDBStatistics();
-  options.row_cache = NewLRUCache(8192);
+  LRUCacheOptions cache_options;
+  cache_options.capacity = 8192;
+  options.row_cache = cache_options.MakeSharedGeneralCache();
+  // BEGIN check that Cache classes as aliases of each other.
+  // Currently, GeneralCache and BlockCache are aliases for Cache.
+  // This is expected to change (carefully, intentionally)
+  std::shared_ptr<GeneralCache> general_cache = options.row_cache;
+  std::shared_ptr<Cache> cache = general_cache;
+  std::shared_ptr<BlockCache> block_cache = general_cache;
+  general_cache = cache;
+  block_cache = cache;
+  general_cache = block_cache;
+  cache = block_cache;
+  // END check that Cache classes as aliases of each other.
  DestroyAndReopen(options);

  ASSERT_OK(Put("foo", "bar"));
@@ -3044,7 +3044,7 @@ TEST_F(DBTest2, PausingManualCompaction1) {
                  .IsManualCompactionPaused());

  // Wait for compactions to get scheduled and stopped
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  // Get file names after compaction is stopped
  files_meta.clear();
@@ -3064,7 +3064,7 @@ TEST_F(DBTest2, PausingManualCompaction1) {
                                 files_before_compact, 0)
                  .IsManualCompactionPaused());
  // Wait for manual compaction to get scheduled and finish
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  files_meta.clear();
  files_after_compact.clear();
@@ -3097,7 +3097,7 @@ TEST_F(DBTest2, PausingManualCompaction2) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  std::vector<LiveFileMetaData> files_meta;
  dbfull()->GetLiveFilesMetaData(&files_meta);
@@ -3139,7 +3139,7 @@ TEST_F(DBTest2, PausingManualCompaction3) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  // As manual compaction disabled, not even reach sync point
  ASSERT_EQ(run_manual_compactions, 0);
  ASSERT_EQ("2,3,4,5,6,7,8", FilesPerLevel());
@@ -3148,7 +3148,7 @@ TEST_F(DBTest2, PausingManualCompaction3) {
      "CompactionJob::Run():PausingManualCompaction:1");
  dbfull()->EnableManualCompaction();
  ASSERT_OK(dbfull()->CompactRange(compact_options, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,0,2", FilesPerLevel());

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
@@ -3203,14 +3203,14 @@ TEST_F(DBTest2, PausingManualCompaction4) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ(run_manual_compactions, 1);
  ASSERT_EQ("2,3,4,5,6,7,8", FilesPerLevel());

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearCallBack(
      "CompactionJob::Run():PausingManualCompaction:2");
  ASSERT_OK(dbfull()->CompactRange(compact_options, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,0,2", FilesPerLevel());

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
@@ -3262,7 +3262,7 @@ TEST_F(DBTest2, CancelManualCompaction1) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  // Since compactions are disabled, we shouldn't start compacting.
  // E.g. we should call the compaction function exactly one time.
@@ -3286,7 +3286,7 @@ TEST_F(DBTest2, CancelManualCompaction1) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_EQ(compactions_run, 3);

@@ -3298,7 +3298,7 @@ TEST_F(DBTest2, CancelManualCompaction1) {
  // Compactions should work again if we re-enable them..
  compact_options.canceled->store(false, std::memory_order_relaxed);
  ASSERT_OK(dbfull()->CompactRange(compact_options, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,0,2", FilesPerLevel());

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
@@ -3362,7 +3362,7 @@ TEST_F(DBTest2, CancelManualCompaction2) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  // NOTE: as we set compact_options.max_subcompacitons = 1, and store true to
  // the canceled variable from the single compacting thread (via callback),
@@ -3380,7 +3380,7 @@ TEST_F(DBTest2, CancelManualCompaction2) {
  // Compactions should work again if we re-enable them..
  compact_options.canceled->store(false, std::memory_order_relaxed);
  ASSERT_OK(dbfull()->CompactRange(compact_options, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,0,0,0,0,0,2", FilesPerLevel());

  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
@@ -3455,7 +3455,7 @@ TEST_F(DBTest2, CancelManualCompactionWithListener) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_GT(listener->num_compaction_started_, 0);
  ASSERT_EQ(listener->num_compaction_started_, listener->num_compaction_ended_);
@@ -3470,7 +3470,7 @@ TEST_F(DBTest2, CancelManualCompactionWithListener) {
  ASSERT_TRUE(dbfull()
                  ->CompactRange(compact_options, nullptr, nullptr)
                  .IsManualCompactionPaused());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_EQ(listener->num_compaction_started_, 0);
  ASSERT_EQ(listener->num_compaction_started_, listener->num_compaction_ended_);
@@ -3492,7 +3492,7 @@ TEST_F(DBTest2, CancelManualCompactionWithListener) {

  compact_options.canceled->store(false, std::memory_order_release);
  ASSERT_OK(dbfull()->CompactRange(compact_options, nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  ASSERT_GT(listener->num_compaction_started_, 0);
  ASSERT_EQ(listener->num_compaction_started_, listener->num_compaction_ended_);
@@ -5298,88 +5298,6 @@ TEST_F(DBTest2, DISABLED_IteratorPinnedMemory) {
  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
 }

-TEST_F(DBTest2, TestBBTTailPrefetch) {
-  std::atomic<bool> called(false);
-  size_t expected_lower_bound = 512 * 1024;
-  size_t expected_higher_bound = 512 * 1024;
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
-      "BlockBasedTable::Open::TailPrefetchLen", [&](void* arg) {
-        size_t* prefetch_size = static_cast<size_t*>(arg);
-        EXPECT_LE(expected_lower_bound, *prefetch_size);
-        EXPECT_GE(expected_higher_bound, *prefetch_size);
-        called = true;
-      });
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
-
-  ASSERT_OK(Put("1", "1"));
-  ASSERT_OK(Put("9", "1"));
-  ASSERT_OK(Flush());
-
-  expected_lower_bound = 0;
-  expected_higher_bound = 8 * 1024;
-
-  ASSERT_OK(Put("1", "1"));
-  ASSERT_OK(Put("9", "1"));
-  ASSERT_OK(Flush());
-
-  ASSERT_OK(Put("1", "1"));
-  ASSERT_OK(Put("9", "1"));
-  ASSERT_OK(Flush());
-
-  // Full compaction to make sure there is no L0 file after the open.
-  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-
-  ASSERT_TRUE(called.load());
-  called = false;
-
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
-
-  std::atomic<bool> first_call(true);
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->SetCallBack(
-      "BlockBasedTable::Open::TailPrefetchLen", [&](void* arg) {
-        size_t* prefetch_size = static_cast<size_t*>(arg);
-        if (first_call) {
-          EXPECT_EQ(4 * 1024, *prefetch_size);
-          first_call = false;
-        } else {
-          EXPECT_GE(4 * 1024, *prefetch_size);
-        }
-        called = true;
-      });
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->EnableProcessing();
-
-  Options options = CurrentOptions();
-  options.max_file_opening_threads = 1;  // one thread
-  BlockBasedTableOptions table_options;
-  table_options.cache_index_and_filter_blocks = true;
-  options.table_factory.reset(NewBlockBasedTableFactory(table_options));
-  options.max_open_files = -1;
-  Reopen(options);
-
-  ASSERT_OK(Put("1", "1"));
-  ASSERT_OK(Put("9", "1"));
-  ASSERT_OK(Flush());
-
-  ASSERT_OK(Put("1", "1"));
-  ASSERT_OK(Put("9", "1"));
-  ASSERT_OK(Flush());
-
-  ASSERT_TRUE(called.load());
-  called = false;
-
-  // Parallel loading SST files
-  options.max_file_opening_threads = 16;
-  Reopen(options);
-
-  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-
-  ASSERT_TRUE(called.load());
-
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->DisableProcessing();
-  ROCKSDB_NAMESPACE::SyncPoint::GetInstance()->ClearAllCallBacks();
-}
-
 TEST_F(DBTest2, TestGetColumnFamilyHandleUnlocked) {
  // Setup sync point dependency to reproduce the race condition of
  // DBImpl::GetColumnFamilyHandleUnlocked
@@ -5905,7 +5823,7 @@ TEST_F(DBTest2, SameSmallestInSameLevel) {
  ASSERT_OK(Flush());
  ASSERT_OK(db_->Merge(WriteOptions(), "key", "8"));
  ASSERT_OK(Flush());
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());
  ASSERT_EQ("0,4,1", FilesPerLevel());

  ASSERT_EQ("2,3,4,5,6,7,8", Get("key"));
@@ -6074,17 +5992,15 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      iterator->Seek("xa");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
-      // It's a bug that the counter BLOOM_FILTER_PREFIX_CHECKED is not
-      // correct in this case. So don't check counters in this case.
      if (expect_filter_check) {
-        ASSERT_EQ(0, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(0, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      iterator->Seek("xz");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xz1", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(0, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(0, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }
    }

@@ -6102,7 +6018,7 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(0, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(0, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }
    }

@@ -6116,14 +6032,14 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(0, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(0, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      iterator->Seek("xx0");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xx1", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(1, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }
    }

@@ -6141,21 +6057,21 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(2, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      iterator->Seek("xg");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xx1", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(3, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      iterator->Seek("xz");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xz1", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(4, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      ASSERT_OK(iterator->status());
@@ -6167,14 +6083,14 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(5, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      iterator->Seek("xx0");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xx1", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(6, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }

      ASSERT_OK(iterator->status());
@@ -6188,7 +6104,7 @@ TEST_F(DBTest2, ChangePrefixExtractor) {
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("xb", iterator->key().ToString());
      if (expect_filter_check) {
-        ASSERT_EQ(7, TestGetTickerCount(options, BLOOM_FILTER_PREFIX_CHECKED));
+        EXPECT_EQ(1, PopTicker(options, NON_LAST_LEVEL_SEEK_FILTER_MATCH));
      }
      ASSERT_OK(iterator->status());
    }
@@ -6262,13 +6178,19 @@ TEST_F(DBTest2, AutoPrefixMode1) {
    ro.total_order_seek = false;
    ro.auto_prefix_mode = true;

-    const auto stat = BLOOM_FILTER_PREFIX_CHECKED;
+    const auto hit_stat = options.num_levels == 1
+                              ? LAST_LEVEL_SEEK_FILTER_MATCH
+                              : NON_LAST_LEVEL_SEEK_FILTER_MATCH;
+    const auto miss_stat = options.num_levels == 1
+                               ? LAST_LEVEL_SEEK_FILTERED
+                               : NON_LAST_LEVEL_SEEK_FILTERED;
    {
      std::unique_ptr<Iterator> iterator(db_->NewIterator(ro));
      iterator->Seek("b1");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("x1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }

@@ -6280,7 +6202,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      std::unique_ptr<Iterator> iterator(db_->NewIterator(ro));
      iterator->Seek("b1");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }

@@ -6290,7 +6213,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      iterator->Seek("b1");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("x1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }

@@ -6299,7 +6223,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      std::unique_ptr<Iterator> iterator(db_->NewIterator(ro));
      iterator->Seek("b1");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }

@@ -6308,7 +6233,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      std::unique_ptr<Iterator> iterator(db_->NewIterator(ro));
      iterator->Seek("b1");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }

@@ -6319,25 +6245,29 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      ub = "b9";
      iterator->Seek("b1");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());

      ub = "z";
      iterator->Seek("b1");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("x1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "c";
      iterator->Seek("b1");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));

      ub = "b9";
      iterator->SeekForPrev("b1");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("a1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "zz";
      iterator->SeekToLast();
@@ -6369,26 +6299,30 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      ub = "b1";
      iterator->Seek("b9");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());

      ub = "b1";
      iterator->Seek("z");
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("y1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "b1";
      iterator->Seek("c");
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "b";
      iterator->Seek("c9");
      ASSERT_FALSE(iterator->Valid());
      // Fails if ReverseBytewiseComparator::IsSameLengthImmediateSuccessor
      // is "correctly" implemented.
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "a";
      iterator->Seek("b9");
@@ -6396,7 +6330,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      // is "correctly" implemented.
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("a1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "b";
      iterator->Seek("a");
@@ -6404,7 +6339,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      // Fails if ReverseBytewiseComparator::IsSameLengthImmediateSuccessor
      // matches BytewiseComparator::IsSameLengthImmediateSuccessor. Upper
      // comparing before seek key prevents a real bug from surfacing.
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "b1";
      iterator->SeekForPrev("b9");
@@ -6412,7 +6348,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      // Fails if ReverseBytewiseComparator::IsSameLengthImmediateSuccessor
      // is "correctly" implemented.
      ASSERT_EQ("x1", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));

      ub = "a";
      iterator->SeekToLast();
@@ -6454,7 +6391,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      std::unique_ptr<Iterator> iterator(db_->NewIterator(ro));
      iterator->Seek(Slice(a_end_stuff, 2));
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());

      // test, cannot be validly optimized with auto_prefix_mode
@@ -6464,7 +6402,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      iterator->Seek(Slice(a_end_stuff, 2));
      // !!! BUG !!! See "BUG" section of auto_prefix_mode.
      ASSERT_FALSE(iterator->Valid());
-      EXPECT_EQ(1, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(1, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());

      // To prove that is the wrong result, now use total order seek
@@ -6475,7 +6414,8 @@ TEST_F(DBTest2, AutoPrefixMode1) {
      iterator->Seek(Slice(a_end_stuff, 2));
      ASSERT_TRUE(iterator->Valid());
      ASSERT_EQ("b", iterator->key().ToString());
-      EXPECT_EQ(0, TestGetAndResetTickerCount(options, stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, hit_stat));
+      EXPECT_EQ(0, TestGetAndResetTickerCount(options, miss_stat));
      ASSERT_OK(iterator->status());
    }
  } while (ChangeOptions(kSkipPlainTable));
@@ -1074,6 +1074,24 @@ size_t DBTestBase::TotalLiveFiles(int cf) {
  return num_files;
 }

+size_t DBTestBase::TotalLiveFilesAtPath(int cf, const std::string& path) {
+  ColumnFamilyMetaData cf_meta;
+  if (cf == 0) {
+    db_->GetColumnFamilyMetaData(&cf_meta);
+  } else {
+    db_->GetColumnFamilyMetaData(handles_[cf], &cf_meta);
+  }
+  size_t num_files = 0;
+  for (auto& level : cf_meta.levels) {
+    for (auto& f : level.files) {
+      if (f.directory == path) {
+        num_files++;
+      }
+    }
+  }
+  return num_files;
+}
+
 size_t DBTestBase::CountLiveFiles() {
  std::vector<LiveFileMetaData> metadata;
  db_->GetLiveFilesMetaData(&metadata);
@@ -1192,6 +1192,8 @@ class DBTestBase : public testing::Test {

  size_t TotalLiveFiles(int cf = 0);

+  size_t TotalLiveFilesAtPath(int cf, const std::string& path);
+
  size_t CountLiveFiles();

  int NumTableFilesAtLevel(int level, int cf = 0);
@@ -1316,6 +1318,10 @@ class DBTestBase : public testing::Test {
                                      Tickers ticker_type) {
    return options.statistics->getAndResetTickerCount(ticker_type);
  }
+  // Short name for TestGetAndResetTickerCount
+  uint64_t PopTicker(const Options& options, Tickers ticker_type) {
+    return options.statistics->getAndResetTickerCount(ticker_type);
+  }

  // Note: reverting this setting within the same test run is not yet
  // supported
@@ -1850,31 +1850,31 @@ TEST_P(DBTestUniversalManualCompactionOutputPathId,
  compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
  ASSERT_OK(db_->CompactRange(compact_options, handles_[1], nullptr, nullptr));
  ASSERT_EQ(1, TotalLiveFiles(1));
-  ASSERT_EQ(0, GetSstFileCount(options.db_paths[0].path));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
+  ASSERT_EQ(0, TotalLiveFilesAtPath(1, options.db_paths[0].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[1].path));

  ReopenWithColumnFamilies({kDefaultColumnFamilyName, "pikachu"}, options);
  ASSERT_EQ(1, TotalLiveFiles(1));
-  ASSERT_EQ(0, GetSstFileCount(options.db_paths[0].path));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
+  ASSERT_EQ(0, TotalLiveFilesAtPath(1, options.db_paths[0].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[1].path));

  MakeTables(1, "p", "q", 1);
  ASSERT_EQ(2, TotalLiveFiles(1));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[0].path));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[0].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[1].path));

  ReopenWithColumnFamilies({kDefaultColumnFamilyName, "pikachu"}, options);
  ASSERT_EQ(2, TotalLiveFiles(1));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[0].path));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[1].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[0].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[1].path));

  // Full compaction to DB path 0
  compact_options.target_path_id = 0;
  compact_options.exclusive_manual_compaction = exclusive_manual_compaction_;
  ASSERT_OK(db_->CompactRange(compact_options, handles_[1], nullptr, nullptr));
  ASSERT_EQ(1, TotalLiveFiles(1));
-  ASSERT_EQ(1, GetSstFileCount(options.db_paths[0].path));
-  ASSERT_EQ(0, GetSstFileCount(options.db_paths[1].path));
+  ASSERT_EQ(1, TotalLiveFilesAtPath(1, options.db_paths[0].path));
+  ASSERT_EQ(0, TotalLiveFilesAtPath(1, options.db_paths[1].path));

  // Fail when compacting to an invalid path ID
  compact_options.target_path_id = 2;
@@ -2399,6 +2399,31 @@ TEST_F(DBWALTest, GetCompressedWalsAfterSync) {
  Status s = dbfull()->GetSortedWalFiles(wals);
  ASSERT_OK(s);
 }
+
+TEST_F(DBWALTest, EmptyWalReopenTest) {
+  Options options = CurrentOptions();
+  options.env = env_;
+  CreateAndReopenWithCF({"pikachu"}, options);
+
+  // make sure we can re-open it.
+  ASSERT_OK(TryReopenWithColumnFamilies({"default", "pikachu"}, options));
+
+  {
+    std::vector<std::string> files;
+    int num_wal_files = 0;
+    ASSERT_OK(env_->GetChildren(dbname_, &files));
+    for (const auto& file : files) {
+      uint64_t number = 0;
+      FileType type = kWalFile;
+      if (ParseFileName(file, &number, &type) && type == kWalFile) {
+        num_wal_files++;
+      }
+    }
+
+    ASSERT_EQ(num_wal_files, 1);
+  }
+}
+
 }  // namespace ROCKSDB_NAMESPACE

 int main(int argc, char** argv) {
@@ -1578,7 +1578,7 @@ TEST_F(DBErrorHandlingFSTest, MultiDBCompactionError) {
  }

  for (auto i = 0; i < kNumDbInstances; ++i) {
-    Status s = static_cast<DBImpl*>(db[i])->TEST_WaitForCompact(true);
+    Status s = static_cast<DBImpl*>(db[i])->TEST_WaitForCompact();
    ASSERT_EQ(s.severity(), Status::Severity::kSoftError);
    fault_fs[i]->SetFilesystemActive(true);
  }
@@ -1587,7 +1587,7 @@ TEST_F(DBErrorHandlingFSTest, MultiDBCompactionError) {
  for (auto i = 0; i < kNumDbInstances; ++i) {
    std::string prop;
    ASSERT_EQ(listener[i]->WaitForRecovery(5000000), true);
-    ASSERT_OK(static_cast<DBImpl*>(db[i])->TEST_WaitForCompact(true));
+    ASSERT_OK(static_cast<DBImpl*>(db[i])->TEST_WaitForCompact());
    EXPECT_TRUE(db[i]->GetProperty(
        "rocksdb.num-files-at-level" + std::to_string(0), &prop));
    EXPECT_EQ(atoi(prop.c_str()), 0);
@@ -1701,7 +1701,7 @@ TEST_F(DBErrorHandlingFSTest, MultiDBVariousErrors) {
  }

  for (auto i = 0; i < kNumDbInstances; ++i) {
-    Status s = static_cast<DBImpl*>(db[i])->TEST_WaitForCompact(true);
+    Status s = static_cast<DBImpl*>(db[i])->TEST_WaitForCompact();
    switch (i) {
      case 0:
        ASSERT_EQ(s.severity(), Status::Severity::kSoftError);
@@ -1723,7 +1723,7 @@ TEST_F(DBErrorHandlingFSTest, MultiDBVariousErrors) {
      ASSERT_EQ(listener[i]->WaitForRecovery(5000000), true);
    }
    if (i == 1) {
-      ASSERT_OK(static_cast<DBImpl*>(db[i])->TEST_WaitForCompact(true));
+      ASSERT_OK(static_cast<DBImpl*>(db[i])->TEST_WaitForCompact());
    }
    EXPECT_TRUE(db[i]->GetProperty(
        "rocksdb.num-files-at-level" + std::to_string(0), &prop));
@@ -102,7 +102,7 @@ Status UpdateManifestForFilesState(
                  lf->oldest_blob_file_number, lf->oldest_ancester_time,
                  lf->file_creation_time, lf->epoch_number, lf->file_checksum,
                  lf->file_checksum_func_name, lf->unique_id,
-                  lf->compensated_range_deletion_size);
+                  lf->compensated_range_deletion_size, lf->tail_size);
            }
          }
        } else {
@@ -464,6 +464,16 @@ Status ExternalSstFileIngestionJob::Run() {
      current_time = oldest_ancester_time =
          static_cast<uint64_t>(temp_current_time);
    }
+    uint64_t tail_size = 0;
+    bool contain_no_data_blocks = f.table_properties.num_entries > 0 &&
+                                  (f.table_properties.num_entries ==
+                                   f.table_properties.num_range_deletions);
+    if (f.table_properties.tail_start_offset > 0 || contain_no_data_blocks) {
+      uint64_t file_size = f.fd.GetFileSize();
+      assert(f.table_properties.tail_start_offset <= file_size);
+      tail_size = file_size - f.table_properties.tail_start_offset;
+    }
+
    FileMetaData f_metadata(
        f.fd.GetNumber(), f.fd.GetPathId(), f.fd.GetFileSize(),
        f.smallest_internal_key, f.largest_internal_key, f.assigned_seqno,
@@ -472,7 +482,7 @@ Status ExternalSstFileIngestionJob::Run() {
        ingestion_options_.ingest_behind
            ? kReservedEpochNumberForFileIngestedBehind
            : cfd_->NewEpochNumber(),
-        f.file_checksum, f.file_checksum_func_name, f.unique_id, 0);
+        f.file_checksum, f.file_checksum_func_name, f.unique_id, 0, tail_size);
    f_metadata.temperature = f.file_temperature;
    edit_.AddFile(f.picked_level, f_metadata);
  }
@@ -678,6 +688,7 @@ Status ExternalSstFileIngestionJob::GetIngestedFileInfo(
      TableReaderOptions(
          *cfd_->ioptions(), sv->mutable_cf_options.prefix_extractor,
          env_options_, cfd_->internal_comparator(),
+          sv->mutable_cf_options.block_protection_bytes_per_key,
          /*skip_filters*/ false, /*immortal*/ false,
          /*force_direct_prefetch*/ false, /*level*/ -1,
          /*block_cache_tracer*/ nullptr,
@@ -1289,7 +1289,7 @@ TEST_F(ExternalSSTFileTest, IngestNonExistingFile) {
  ASSERT_OK(Flush());

  ASSERT_OK(db_->CompactRange(CompactRangeOptions(), nullptr, nullptr));
-  ASSERT_OK(dbfull()->TEST_WaitForCompact(true));
+  ASSERT_OK(dbfull()->TEST_WaitForCompact());

  // After full compaction, there should be only 1 file.
  std::vector<std::string> files;
@@ -2857,4 +2857,3 @@ int main(int argc, char** argv) {
  ::testing::InitGoogleTest(&argc, argv);
  return RUN_ALL_TESTS();
 }
-
@@ -1007,7 +1007,8 @@ Status FlushJob::WriteLevel0Table() {
                   meta_.oldest_blob_file_number, meta_.oldest_ancester_time,
                   meta_.file_creation_time, meta_.epoch_number,
                   meta_.file_checksum, meta_.file_checksum_func_name,
-                   meta_.unique_id, meta_.compensated_range_deletion_size);
+                   meta_.unique_id, meta_.compensated_range_deletion_size,
+                   meta_.tail_size);
    edit_->SetBlobFileAdditions(std::move(blob_file_additions));
  }
  // Piggyback FlushJobInfo on the first first flushed memtable.
@@ -36,7 +36,7 @@ class ForwardLevelIterator : public InternalIterator {
      const ColumnFamilyData* const cfd, const ReadOptions& read_options,
      const std::vector<FileMetaData*>& files,
      const std::shared_ptr<const SliceTransform>& prefix_extractor,
-      bool allow_unprepared_value)
+      bool allow_unprepared_value, uint8_t block_protection_bytes_per_key)
      : cfd_(cfd),
        read_options_(read_options),
        files_(files),
@@ -45,7 +45,8 @@ class ForwardLevelIterator : public InternalIterator {
        file_iter_(nullptr),
        pinned_iters_mgr_(nullptr),
        prefix_extractor_(prefix_extractor),
-        allow_unprepared_value_(allow_unprepared_value) {
+        allow_unprepared_value_(allow_unprepared_value),
+        block_protection_bytes_per_key_(block_protection_bytes_per_key) {
    status_.PermitUncheckedError();  // Allow uninitialized status through
  }

@@ -87,7 +88,8 @@ class ForwardLevelIterator : public InternalIterator {
        /*arena=*/nullptr, /*skip_filters=*/false, /*level=*/-1,
        /*max_file_size_for_l0_meta_pin=*/0,
        /*smallest_compaction_key=*/nullptr,
-        /*largest_compaction_key=*/nullptr, allow_unprepared_value_);
+        /*largest_compaction_key=*/nullptr, allow_unprepared_value_,
+        block_protection_bytes_per_key_);
    file_iter_->SetPinnedItersMgr(pinned_iters_mgr_);
    valid_ = false;
    if (!range_del_agg.IsEmpty()) {
@@ -211,6 +213,7 @@ class ForwardLevelIterator : public InternalIterator {
  // Kept alive by ForwardIterator::sv_->mutable_cf_options
  const std::shared_ptr<const SliceTransform>& prefix_extractor_;
  const bool allow_unprepared_value_;
+  const uint8_t block_protection_bytes_per_key_;
 };

 ForwardIterator::ForwardIterator(DBImpl* db, const ReadOptions& read_options,
@@ -738,7 +741,8 @@ void ForwardIterator::RebuildIterators(bool refresh_sv) {
        /*skip_filters=*/false, /*level=*/-1,
        MaxFileSizeForL0MetaPin(sv_->mutable_cf_options),
        /*smallest_compaction_key=*/nullptr,
-        /*largest_compaction_key=*/nullptr, allow_unprepared_value_));
+        /*largest_compaction_key=*/nullptr, allow_unprepared_value_,
+        sv_->mutable_cf_options.block_protection_bytes_per_key));
  }
  BuildLevelIterators(vstorage, sv_);
  current_ = nullptr;
@@ -819,7 +823,8 @@ void ForwardIterator::RenewIterators() {
        /*skip_filters=*/false, /*level=*/-1,
        MaxFileSizeForL0MetaPin(svnew->mutable_cf_options),
        /*smallest_compaction_key=*/nullptr,
-        /*largest_compaction_key=*/nullptr, allow_unprepared_value_));
+        /*largest_compaction_key=*/nullptr, allow_unprepared_value_,
+        svnew->mutable_cf_options.block_protection_bytes_per_key));
  }

  for (auto* f : l0_iters_) {
@@ -863,7 +868,8 @@ void ForwardIterator::BuildLevelIterators(const VersionStorageInfo* vstorage,
    } else {
      level_iters_.push_back(new ForwardLevelIterator(
          cfd_, read_options_, level_files,
-          sv->mutable_cf_options.prefix_extractor, allow_unprepared_value_));
+          sv->mutable_cf_options.prefix_extractor, allow_unprepared_value_,
+          sv->mutable_cf_options.block_protection_bytes_per_key));
    }
  }
 }
@@ -885,7 +891,8 @@ void ForwardIterator::ResetIncompleteIterators() {
        /*skip_filters=*/false, /*level=*/-1,
        MaxFileSizeForL0MetaPin(sv_->mutable_cf_options),
        /*smallest_compaction_key=*/nullptr,
-        /*largest_compaction_key=*/nullptr, allow_unprepared_value_);
+        /*largest_compaction_key=*/nullptr, allow_unprepared_value_,
+        sv_->mutable_cf_options.block_protection_bytes_per_key);
    l0_iters_[i]->SetPinnedItersMgr(pinned_iters_mgr_);
  }

@@ -138,6 +138,16 @@ Status ImportColumnFamilyJob::Run() {
    const auto& f = files_to_import_[i];
    const auto& file_metadata = metadata_[i];

+    uint64_t tail_size = 0;
+    bool contain_no_data_blocks = f.table_properties.num_entries > 0 &&
+                                  (f.table_properties.num_entries ==
+                                   f.table_properties.num_range_deletions);
+    if (f.table_properties.tail_start_offset > 0 || contain_no_data_blocks) {
+      uint64_t file_size = f.fd.GetFileSize();
+      assert(f.table_properties.tail_start_offset <= file_size);
+      tail_size = file_size - f.table_properties.tail_start_offset;
+    }
+
    VersionEdit dummy_version_edit;
    dummy_version_edit.AddFile(
        file_metadata.level, f.fd.GetNumber(), f.fd.GetPathId(),
@@ -145,7 +155,7 @@ Status ImportColumnFamilyJob::Run() {
        file_metadata.smallest_seqno, file_metadata.largest_seqno, false,
        file_metadata.temperature, kInvalidBlobFileNumber, oldest_ancester_time,
        current_time, file_metadata.epoch_number, kUnknownFileChecksum,
-        kUnknownFileChecksumFuncName, f.unique_id, 0);
+        kUnknownFileChecksumFuncName, f.unique_id, 0, tail_size);
    s = dummy_version_builder.Apply(&dummy_version_edit);
  }
  if (s.ok()) {
@@ -250,6 +260,7 @@ Status ImportColumnFamilyJob::GetIngestedFileInfo(
      TableReaderOptions(
          *cfd_->ioptions(), sv->mutable_cf_options.prefix_extractor,
          env_options_, cfd_->internal_comparator(),
+          sv->mutable_cf_options.block_protection_bytes_per_key,
          /*skip_filters*/ false, /*immortal*/ false,
          /*force_direct_prefetch*/ false, /*level*/ -1,
          /*block_cache_tracer*/ nullptr,
@@ -700,6 +700,7 @@ void InternalStats::CacheEntryRoleStats::BeginCollection(
  cache_usage = cache->GetUsage();
  table_size = cache->GetTableAddressCount();
  occupancy = cache->GetOccupancyCount();
+  hash_seed = cache->GetHashSeed();
 }

 void InternalStats::CacheEntryRoleStats::EndCollection(
@@ -724,7 +725,7 @@ std::string InternalStats::CacheEntryRoleStats::ToString(
  std::ostringstream str;
  str << "Block cache " << cache_id
      << " capacity: " << BytesToHumanString(cache_capacity)
-      << " usage: " << BytesToHumanString(cache_usage)
+      << " seed: " << hash_seed << " usage: " << BytesToHumanString(cache_usage)
      << " table_size: " << table_size << " occupancy: " << occupancy
      << " collections: " << collection_count
      << " last_copies: " << copies_of_last_collection
@@ -1687,8 +1688,16 @@ void InternalStats::DumpDBStatsWriteStall(std::string* value) {
  std::ostringstream str;
  str << "Write Stall (count): ";

-  for (const auto& name_and_stat : write_stall_stats_map) {
-    str << name_and_stat.first << ": " << name_and_stat.second << ", ";
+  for (auto write_stall_stats_map_iter = write_stall_stats_map.begin();
+       write_stall_stats_map_iter != write_stall_stats_map.end();
+       write_stall_stats_map_iter++) {
+    const auto& name_and_stat = *write_stall_stats_map_iter;
+    str << name_and_stat.first << ": " << name_and_stat.second;
+    if (std::next(write_stall_stats_map_iter) == write_stall_stats_map.end()) {
+      str << "\n";
+    } else {
+      str << ", ";
+    }
  }
  *value = str.str();
 }
@@ -1870,8 +1879,16 @@ void InternalStats::DumpCFStatsWriteStall(std::string* value,
  std::ostringstream str;
  str << "Write Stall (count): ";

-  for (const auto& name_and_stat : write_stall_stats_map) {
-    str << name_and_stat.first << ": " << name_and_stat.second << ", ";
+  for (auto write_stall_stats_map_iter = write_stall_stats_map.begin();
+       write_stall_stats_map_iter != write_stall_stats_map.end();
+       write_stall_stats_map_iter++) {
+    const auto& name_and_stat = *write_stall_stats_map_iter;
+    str << name_and_stat.first << ": " << name_and_stat.second;
+    if (std::next(write_stall_stats_map_iter) == write_stall_stats_map.end()) {
+      str << "\n";
+    } else {
+      str << ", ";
+    }
  }

  if (total_stall_count) {
@@ -478,6 +478,7 @@ class InternalStats {
    uint32_t copies_of_last_collection = 0;
    uint64_t last_start_time_micros_ = 0;
    uint64_t last_end_time_micros_ = 0;
+    uint32_t hash_seed = 0;

    void Clear() {
      // Wipe everything except collection_count
@@ -46,6 +46,8 @@ template <typename T>
 class ProtectionInfoKVOC;
 template <typename T>
 class ProtectionInfoKVOS;
+template <typename T>
+class ProtectionInfoKV;

 // Aliases for 64-bit protection infos.
 using ProtectionInfo64 = ProtectionInfo<uint64_t>;
@@ -64,13 +66,13 @@ class ProtectionInfo {
  ProtectionInfoKVO<T> ProtectKVO(const SliceParts& key,
                                  const SliceParts& value,
                                  ValueType op_type) const;
-
-  T GetVal() const { return val_; }
+  ProtectionInfoKV<T> ProtectKV(const Slice& key, const Slice& value) const;

 private:
  friend class ProtectionInfoKVO<T>;
  friend class ProtectionInfoKVOS<T>;
  friend class ProtectionInfoKVOC<T>;
+  friend class ProtectionInfoKV<T>;

  // Each field is hashed with an independent value so we can catch fields being
  // swapped. Per the `NPHash64()` docs, using consecutive seeds is a pitfall,
@@ -89,8 +91,47 @@ class ProtectionInfo {
    static_assert(sizeof(ProtectionInfo<T>) == sizeof(T), "");
  }

+  T GetVal() const { return val_; }
  void SetVal(T val) { val_ = val; }

+  void Encode(uint8_t len, char* dst) const {
+    assert(sizeof(val_) >= len);
+    switch (len) {
+      case 1:
+        dst[0] = static_cast<uint8_t>(val_);
+        break;
+      case 2:
+        EncodeFixed16(dst, static_cast<uint16_t>(val_));
+        break;
+      case 4:
+        EncodeFixed32(dst, static_cast<uint32_t>(val_));
+        break;
+      case 8:
+        EncodeFixed64(dst, static_cast<uint64_t>(val_));
+        break;
+      default:
+        assert(false);
+    }
+  }
+
+  bool Verify(uint8_t len, const char* checksum_ptr) const {
+    assert(sizeof(val_) >= len);
+    switch (len) {
+      case 1:
+        return static_cast<uint8_t>(checksum_ptr[0]) ==
+               static_cast<uint8_t>(val_);
+      case 2:
+        return DecodeFixed16(checksum_ptr) == static_cast<uint16_t>(val_);
+      case 4:
+        return DecodeFixed32(checksum_ptr) == static_cast<uint32_t>(val_);
+      case 8:
+        return DecodeFixed64(checksum_ptr) == static_cast<uint64_t>(val_);
+      default:
+        assert(false);
+        return false;
+    }
+  }
+
  T val_ = 0;
 };

@@ -113,7 +154,14 @@ class ProtectionInfoKVO {
  void UpdateV(const SliceParts& old_value, const SliceParts& new_value);
  void UpdateO(ValueType old_op_type, ValueType new_op_type);

-  T GetVal() const { return info_.GetVal(); }
+  // Encode this protection info into `len` bytes and stores them in `dst`.
+  void Encode(uint8_t len, char* dst) const { info_.Encode(len, dst); }
+  // Verify this protection info against the protection info encoded by Encode()
+  // at the first `len` bytes of `checksum_ptr`.
+  // Returns true iff the verification is successful.
+  bool Verify(uint8_t len, const char* checksum_ptr) const {
+    return info_.Verify(len, checksum_ptr);
+  }

 private:
  friend class ProtectionInfo<T>;
@@ -124,6 +172,7 @@ class ProtectionInfoKVO {
    static_assert(sizeof(ProtectionInfoKVO<T>) == sizeof(T), "");
  }

+  T GetVal() const { return info_.GetVal(); }
  void SetVal(T val) { info_.SetVal(val); }

  ProtectionInfo<T> info_;
@@ -154,7 +203,10 @@ class ProtectionInfoKVOC {
  void UpdateC(ColumnFamilyId old_column_family_id,
               ColumnFamilyId new_column_family_id);

-  T GetVal() const { return kvo_.GetVal(); }
+  void Encode(uint8_t len, char* dst) const { kvo_.Encode(len, dst); }
+  bool Verify(uint8_t len, const char* checksum_ptr) const {
+    return kvo_.Verify(len, checksum_ptr);
+  }

 private:
  friend class ProtectionInfoKVO<T>;
@@ -163,6 +215,7 @@ class ProtectionInfoKVOC {
    static_assert(sizeof(ProtectionInfoKVOC<T>) == sizeof(T), "");
  }

+  T GetVal() const { return kvo_.GetVal(); }
  void SetVal(T val) { kvo_.SetVal(val); }

  ProtectionInfoKVO<T> kvo_;
@@ -193,7 +246,10 @@ class ProtectionInfoKVOS {
  void UpdateS(SequenceNumber old_sequence_number,
               SequenceNumber new_sequence_number);

-  T GetVal() const { return kvo_.GetVal(); }
+  void Encode(uint8_t len, char* dst) const { kvo_.Encode(len, dst); }
+  bool Verify(uint8_t len, const char* checksum_ptr) const {
+    return kvo_.Verify(len, checksum_ptr);
+  }

 private:
  friend class ProtectionInfoKVO<T>;
@@ -202,11 +258,32 @@ class ProtectionInfoKVOS {
    static_assert(sizeof(ProtectionInfoKVOS<T>) == sizeof(T), "");
  }

+  T GetVal() const { return kvo_.GetVal(); }
  void SetVal(T val) { kvo_.SetVal(val); }

  ProtectionInfoKVO<T> kvo_;
 };

+template <typename T>
+class ProtectionInfoKV {
+ public:
+  ProtectionInfoKV() = default;
+
+  void Encode(uint8_t len, char* dst) const { info_.Encode(len, dst); }
+  bool Verify(uint8_t len, const char* checksum_ptr) const {
+    return info_.Verify(len, checksum_ptr);
+  }
+
+ private:
+  friend class ProtectionInfo<T>;
+
+  explicit ProtectionInfoKV(T val) : info_(val) {
+    static_assert(sizeof(ProtectionInfoKV<T>) == sizeof(T));
+  }
+
+  ProtectionInfo<T> info_;
+};
+
 template <typename T>
 Status ProtectionInfo<T>::GetStatus() const {
  if (val_ != 0) {
@@ -244,6 +321,16 @@ ProtectionInfoKVO<T> ProtectionInfo<T>::ProtectKVO(const SliceParts& key,
  return ProtectionInfoKVO<T>(val);
 }

+template <typename T>
+ProtectionInfoKV<T> ProtectionInfo<T>::ProtectKV(const Slice& key,
+                                                 const Slice& value) const {
+  T val = GetVal();
+  val = val ^ static_cast<T>(GetSliceNPHash64(key, ProtectionInfo<T>::kSeedK));
+  val =
+      val ^ static_cast<T>(GetSliceNPHash64(value, ProtectionInfo<T>::kSeedV));
+  return ProtectionInfoKV<T>(val);
+}
+
 template <typename T>
 void ProtectionInfoKVO<T>::UpdateK(const Slice& old_key, const Slice& new_key) {
  T val = GetVal();
@@ -394,5 +481,4 @@ void ProtectionInfoKVOS<T>::UpdateS(SequenceNumber old_sequence_number,
                  sizeof(new_sequence_number), ProtectionInfo<T>::kSeedS));
  SetVal(val);
 }
-
 }  // namespace ROCKSDB_NAMESPACE
@@ -12,7 +12,7 @@
 #include "db/version_set.h"
 #include "db/write_batch_internal.h"
 #include "file/filename.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/compaction_filter.h"
 #include "rocksdb/db.h"
@@ -29,7 +29,7 @@
 #include "test_util/testutil.h"
 #include "util/hash.h"
 #include "util/mutexlock.h"
-#include "util/rate_limiter.h"
+#include "util/rate_limiter_impl.h"
 #include "util/string_util.h"
 #include "utilities/merge_operators.h"

@@ -35,17 +35,21 @@ enum RecordType {

  // Compression Type
  kSetCompressionType = 9,
-};
-static const int kMaxRecordType = kSetCompressionType;

-static const unsigned int kBlockSize = 32768;
+  // User-defined timestamp sizes
+  kUserDefinedTimestampSizeType = 10,
+  kRecyclableUserDefinedTimestampSizeType = 11,
+};
+constexpr int kMaxRecordType = kRecyclableUserDefinedTimestampSizeType;
+
+constexpr unsigned int kBlockSize = 32768;

 // Header is checksum (4 bytes), length (2 bytes), type (1 byte)
-static const int kHeaderSize = 4 + 2 + 1;
+constexpr int kHeaderSize = 4 + 2 + 1;

 // Recyclable header is checksum (4 bytes), length (2 bytes), type (1 byte),
 // log number (4 bytes).
-static const int kRecyclableHeaderSize = 4 + 2 + 1 + 4;
+constexpr int kRecyclableHeaderSize = 4 + 2 + 1 + 4;

 }  // namespace log
 }  // namespace ROCKSDB_NAMESPACE
@@ -164,6 +164,54 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
        }
        break;

+      case kSetCompressionType: {
+        if (compression_type_record_read_) {
+          ReportCorruption(fragment.size(),
+                           "read multiple SetCompressionType records");
+        }
+        if (first_record_read_) {
+          ReportCorruption(fragment.size(),
+                           "SetCompressionType not the first record");
+        }
+        prospective_record_offset = physical_record_offset;
+        scratch->clear();
+        last_record_offset_ = prospective_record_offset;
+        CompressionTypeRecord compression_record(kNoCompression);
+        Status s = compression_record.DecodeFrom(&fragment);
+        if (!s.ok()) {
+          ReportCorruption(fragment.size(),
+                           "could not decode SetCompressionType record");
+        } else {
+          InitCompression(compression_record);
+        }
+        break;
+      }
+      case kUserDefinedTimestampSizeType:
+      case kRecyclableUserDefinedTimestampSizeType: {
+        if (in_fragmented_record && !scratch->empty()) {
+          ReportCorruption(
+              scratch->size(),
+              "user-defined timestamp size record interspersed partial record");
+        }
+        prospective_record_offset = physical_record_offset;
+        scratch->clear();
+        last_record_offset_ = prospective_record_offset;
+        UserDefinedTimestampSizeRecord ts_record;
+        Status s = ts_record.DecodeFrom(&fragment);
+        if (!s.ok()) {
+          ReportCorruption(
+              fragment.size(),
+              "could not decode user-defined timestamp size record");
+        } else {
+          s = UpdateRecordedTimestampSize(
+              ts_record.GetUserDefinedTimestampSize());
+          if (!s.ok()) {
+            ReportCorruption(fragment.size(), s.getState());
+          }
+        }
+        break;
+      }
+
      case kBadHeader:
        if (wal_recovery_mode == WALRecoveryMode::kAbsoluteConsistency ||
            wal_recovery_mode == WALRecoveryMode::kPointInTimeRecovery) {
@@ -257,29 +305,6 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch,
        }
        break;

-      case kSetCompressionType: {
-        if (compression_type_record_read_) {
-          ReportCorruption(fragment.size(),
-                           "read multiple SetCompressionType records");
-        }
-        if (first_record_read_) {
-          ReportCorruption(fragment.size(),
-                           "SetCompressionType not the first record");
-        }
-        prospective_record_offset = physical_record_offset;
-        scratch->clear();
-        last_record_offset_ = prospective_record_offset;
-        CompressionTypeRecord compression_record(kNoCompression);
-        Status s = compression_record.DecodeFrom(&fragment);
-        if (!s.ok()) {
-          ReportCorruption(fragment.size(),
-                           "could not decode SetCompressionType record");
-        } else {
-          InitCompression(compression_record);
-        }
-        break;
-      }
-
      default: {
        char buf[40];
        snprintf(buf, sizeof(buf), "unknown record type %u", record_type);
@@ -444,7 +469,8 @@ unsigned int Reader::ReadPhysicalRecord(Slice* result, size_t* drop_size,
    const unsigned int type = header[6];
    const uint32_t length = a | (b << 8);
    int header_size = kHeaderSize;
-    if (type >= kRecyclableFullType && type <= kRecyclableLastType) {
+    if ((type >= kRecyclableFullType && type <= kRecyclableLastType) ||
+        type == kRecyclableUserDefinedTimestampSizeType) {
      if (end_of_buffer_offset_ - buffer_.size() == 0) {
        recycled_ = true;
      }
@@ -500,7 +526,9 @@ unsigned int Reader::ReadPhysicalRecord(Slice* result, size_t* drop_size,

    buffer_.remove_prefix(header_size + length);

-    if (!uncompress_ || type == kSetCompressionType) {
+    if (!uncompress_ || type == kSetCompressionType ||
+        type == kUserDefinedTimestampSizeType ||
+        type == kRecyclableUserDefinedTimestampSizeType) {
      *result = Slice(header + header_size, length);
      return type;
    } else {
@@ -567,6 +595,26 @@ void Reader::InitCompression(const CompressionTypeRecord& compression_record) {
  assert(uncompressed_buffer_);
 }

+Status Reader::UpdateRecordedTimestampSize(
+    const std::vector<std::pair<uint32_t, size_t>>& cf_to_ts_sz) {
+  for (const auto& [cf, ts_sz] : cf_to_ts_sz) {
+    // Zero user-defined timestamp size are not recorded.
+    if (ts_sz == 0) {
+      return Status::Corruption(
+          "User-defined timestamp size record contains zero timestamp size.");
+    }
+    // The user-defined timestamp size record for a column family should not be
+    // updated in the same log file.
+    if (recorded_cf_to_ts_sz_.count(cf) != 0) {
+      return Status::Corruption(
+          "User-defined timestamp size record contains update to "
+          "recorded column family.");
+    }
+    recorded_cf_to_ts_sz_.insert(std::make_pair(cf, ts_sz));
+  }
+  return Status::OK();
+}
+
 bool FragmentBufferedReader::ReadRecord(Slice* record, std::string* scratch,
                                        WALRecoveryMode /*unused*/,
                                        uint64_t* /* checksum */) {
@@ -635,30 +683,6 @@ bool FragmentBufferedReader::ReadRecord(Slice* record, std::string* scratch,
        }
        break;

-      case kBadHeader:
-      case kBadRecord:
-      case kEof:
-      case kOldRecord:
-        if (in_fragmented_record_) {
-          ReportCorruption(fragments_.size(), "error in middle of record");
-          in_fragmented_record_ = false;
-          fragments_.clear();
-        }
-        break;
-
-      case kBadRecordChecksum:
-        if (recycled_) {
-          fragments_.clear();
-          return false;
-        }
-        ReportCorruption(drop_size, "checksum mismatch");
-        if (in_fragmented_record_) {
-          ReportCorruption(fragments_.size(), "error in middle of record");
-          in_fragmented_record_ = false;
-          fragments_.clear();
-        }
-        break;
-
      case kSetCompressionType: {
        if (compression_type_record_read_) {
          ReportCorruption(fragment.size(),
@@ -683,6 +707,57 @@ bool FragmentBufferedReader::ReadRecord(Slice* record, std::string* scratch,
        break;
      }

+      case kUserDefinedTimestampSizeType:
+      case kRecyclableUserDefinedTimestampSizeType: {
+        if (in_fragmented_record_ && !scratch->empty()) {
+          ReportCorruption(
+              scratch->size(),
+              "user-defined timestamp size record interspersed partial record");
+        }
+        fragments_.clear();
+        prospective_record_offset = physical_record_offset;
+        last_record_offset_ = prospective_record_offset;
+        in_fragmented_record_ = false;
+        UserDefinedTimestampSizeRecord ts_record;
+        Status s = ts_record.DecodeFrom(&fragment);
+        if (!s.ok()) {
+          ReportCorruption(
+              fragment.size(),
+              "could not decode user-defined timestamp size record");
+        } else {
+          s = UpdateRecordedTimestampSize(
+              ts_record.GetUserDefinedTimestampSize());
+          if (!s.ok()) {
+            ReportCorruption(fragment.size(), s.getState());
+          }
+        }
+        break;
+      }
+
+      case kBadHeader:
+      case kBadRecord:
+      case kEof:
+      case kOldRecord:
+        if (in_fragmented_record_) {
+          ReportCorruption(fragments_.size(), "error in middle of record");
+          in_fragmented_record_ = false;
+          fragments_.clear();
+        }
+        break;
+
+      case kBadRecordChecksum:
+        if (recycled_) {
+          fragments_.clear();
+          return false;
+        }
+        ReportCorruption(drop_size, "checksum mismatch");
+        if (in_fragmented_record_) {
+          ReportCorruption(fragments_.size(), "error in middle of record");
+          in_fragmented_record_ = false;
+          fragments_.clear();
+        }
+        break;
+
      default: {
        char buf[40];
        snprintf(buf, sizeof(buf), "unknown record type %u",
@@ -770,7 +845,8 @@ bool FragmentBufferedReader::TryReadFragment(
  const unsigned int type = header[6];
  const uint32_t length = a | (b << 8);
  int header_size = kHeaderSize;
-  if (type >= kRecyclableFullType && type <= kRecyclableLastType) {
+  if ((type >= kRecyclableFullType && type <= kRecyclableLastType) ||
+      type == kRecyclableUserDefinedTimestampSizeType) {
    if (end_of_buffer_offset_ - buffer_.size() == 0) {
      recycled_ = true;
    }
@@ -822,7 +898,9 @@ bool FragmentBufferedReader::TryReadFragment(

  buffer_.remove_prefix(header_size + length);

-  if (!uncompress_ || type == kSetCompressionType) {
+  if (!uncompress_ || type == kSetCompressionType ||
+      type == kUserDefinedTimestampSizeType ||
+      type == kRecyclableUserDefinedTimestampSizeType) {
    *fragment = Slice(header + header_size, length);
    *fragment_type_or_err = type;
    return true;
@@ -11,6 +11,8 @@
 #include <stdint.h>

 #include <memory>
+#include <unordered_map>
+#include <vector>

 #include "db/log_format.h"
 #include "file/sequence_file_reader.h"
@@ -18,6 +20,7 @@
 #include "rocksdb/slice.h"
 #include "rocksdb/status.h"
 #include "util/compression.h"
+#include "util/udt_util.h"
 #include "util/xxhash.h"

 namespace ROCKSDB_NAMESPACE {
@@ -74,6 +77,12 @@ class Reader {
                              WALRecoveryMode::kTolerateCorruptedTailRecords,
                          uint64_t* record_checksum = nullptr);

+  // Return the recorded user-defined timestamp size that have been read so
+  // far. This only applies to WAL logs.
+  const std::unordered_map<uint32_t, size_t>& GetRecordedTimestampSize() const {
+    return recorded_cf_to_ts_sz_;
+  }
+
  // Returns the physical offset of the last record returned by ReadRecord.
  //
  // Undefined before the first call to ReadRecord.
@@ -154,6 +163,10 @@ class Reader {
  // Used for stream hashing uncompressed buffer in ReadPhysicalRecord()
  XXH3_state_t* uncompress_hash_state_;

+  // The recorded user-defined timestamp sizes that have been read so far. This
+  // is only for WAL logs.
+  std::unordered_map<uint32_t, size_t> recorded_cf_to_ts_sz_;
+
  // Extend record types with the following special values
  enum {
    kEof = kMaxRecordType + 1,
@@ -190,6 +203,9 @@ class Reader {
  void ReportDrop(size_t bytes, const Status& reason);

  void InitCompression(const CompressionTypeRecord& compression_record);
+
+  Status UpdateRecordedTimestampSize(
+      const std::vector<std::pair<uint32_t, size_t>>& cf_to_ts_sz);
 };

 class FragmentBufferedReader : public Reader {
@@ -45,9 +45,10 @@ static std::string RandomSkewedString(int i, Random* rnd) {
  return BigString(NumberString(i), rnd->Skewed(17));
 }

-// Param type is tuple<int, bool>
+// Param type is tuple<int, bool, CompressionType>
 // get<0>(tuple): non-zero if recycling log, zero if regular log
 // get<1>(tuple): true if allow retry after read EOF, false otherwise
+// get<2>(tuple): type of compression used
 class LogTest
    : public ::testing::TestWithParam<std::tuple<int, bool, CompressionType>> {
 private:
@@ -181,20 +182,30 @@ class LogTest

  Slice* get_reader_contents() { return &reader_contents_; }

-  void Write(const std::string& msg) {
+  void Write(
+      const std::string& msg,
+      const std::unordered_map<uint32_t, size_t>* cf_to_ts_sz = nullptr) {
+    if (cf_to_ts_sz != nullptr && !cf_to_ts_sz->empty()) {
+      ASSERT_OK(writer_->MaybeAddUserDefinedTimestampSizeRecord(*cf_to_ts_sz));
+    }
    ASSERT_OK(writer_->AddRecord(Slice(msg)));
  }

  size_t WrittenBytes() const { return dest_contents().size(); }

-  std::string Read(const WALRecoveryMode wal_recovery_mode =
-                       WALRecoveryMode::kTolerateCorruptedTailRecords) {
+  std::string Read(
+      const WALRecoveryMode wal_recovery_mode =
+          WALRecoveryMode::kTolerateCorruptedTailRecords,
+      std::unordered_map<uint32_t, size_t>* cf_to_ts_sz = nullptr) {
    std::string scratch;
    Slice record;
    bool ret = false;
    uint64_t record_checksum;
    ret = reader_->ReadRecord(&record, &scratch, wal_recovery_mode,
                              &record_checksum);
+    if (cf_to_ts_sz != nullptr) {
+      *cf_to_ts_sz = reader_->GetRecordedTimestampSize();
+    }
    if (ret) {
      if (!allow_retry_read_) {
        // allow_retry_read_ means using FragmentBufferedReader which does not
@@ -257,6 +268,17 @@ class LogTest
      return "OK";
    }
  }
+
+  void CheckRecordAndTimestampSize(
+      std::string record,
+      std::unordered_map<uint32_t, size_t>& expected_ts_sz) {
+    std::unordered_map<uint32_t, size_t> recorded_ts_sz;
+    ASSERT_EQ(record,
+              Read(WALRecoveryMode::
+                       kTolerateCorruptedTailRecords /* wal_recovery_mode */,
+                   &recorded_ts_sz));
+    EXPECT_EQ(expected_ts_sz, recorded_ts_sz);
+  }
 };

 TEST_P(LogTest, Empty) { ASSERT_EQ("EOF", Read()); }
@@ -274,6 +296,43 @@ TEST_P(LogTest, ReadWrite) {
  ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
 }

+TEST_P(LogTest, ReadWriteWithTimestampSize) {
+  std::unordered_map<uint32_t, size_t> ts_sz_one = {
+      {1, sizeof(uint64_t)},
+  };
+  Write("foo", &ts_sz_one);
+  Write("bar");
+  std::unordered_map<uint32_t, size_t> ts_sz_two = {{2, sizeof(char)}};
+  Write("", &ts_sz_two);
+  Write("xxxx");
+
+  CheckRecordAndTimestampSize("foo", ts_sz_one);
+  CheckRecordAndTimestampSize("bar", ts_sz_one);
+  std::unordered_map<uint32_t, size_t> expected_ts_sz_two;
+  // User-defined timestamp size records are accumulated and applied to
+  // subsequent records.
+  expected_ts_sz_two.insert(ts_sz_one.begin(), ts_sz_one.end());
+  expected_ts_sz_two.insert(ts_sz_two.begin(), ts_sz_two.end());
+  CheckRecordAndTimestampSize("", expected_ts_sz_two);
+  CheckRecordAndTimestampSize("xxxx", expected_ts_sz_two);
+  ASSERT_EQ("EOF", Read());
+  ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
+}
+
+TEST_P(LogTest, ReadWriteWithTimestampSizeZeroTimestampIgnored) {
+  std::unordered_map<uint32_t, size_t> ts_sz_one = {{1, sizeof(uint64_t)}};
+  Write("foo", &ts_sz_one);
+  std::unordered_map<uint32_t, size_t> ts_sz_two(ts_sz_one.begin(),
+                                                 ts_sz_one.end());
+  ts_sz_two.insert(std::make_pair(2, 0));
+  Write("bar", &ts_sz_two);
+
+  CheckRecordAndTimestampSize("foo", ts_sz_one);
+  CheckRecordAndTimestampSize("bar", ts_sz_one);
+  ASSERT_EQ("EOF", Read());
+  ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
+}
+
 TEST_P(LogTest, ManyBlocks) {
  for (int i = 0; i < 100000; i++) {
    Write(NumberString(i));
@@ -685,6 +744,39 @@ TEST_P(LogTest, Recycle) {
  ASSERT_EQ("EOF", Read());
 }

+TEST_P(LogTest, RecycleWithTimestampSize) {
+  bool recyclable_log = (std::get<0>(GetParam()) != 0);
+  if (!recyclable_log) {
+    return;  // test is only valid for recycled logs
+  }
+  std::unordered_map<uint32_t, size_t> ts_sz_one = {
+      {1, sizeof(uint32_t)},
+  };
+  Write("foo", &ts_sz_one);
+  Write("bar");
+  Write("baz");
+  Write("bif");
+  Write("blitz");
+  while (get_reader_contents()->size() < log::kBlockSize * 2) {
+    Write("xxxxxxxxxxxxxxxx");
+  }
+  std::unique_ptr<FSWritableFile> sink(
+      new test::OverwritingStringSink(get_reader_contents()));
+  std::unique_ptr<WritableFileWriter> dest_holder(new WritableFileWriter(
+      std::move(sink), "" /* don't care */, FileOptions()));
+  Writer recycle_writer(std::move(dest_holder), 123, true);
+  std::unordered_map<uint32_t, size_t> ts_sz_two = {
+      {2, sizeof(uint64_t)},
+  };
+  ASSERT_OK(recycle_writer.MaybeAddUserDefinedTimestampSizeRecord(ts_sz_two));
+  ASSERT_OK(recycle_writer.AddRecord(Slice("foooo")));
+  ASSERT_OK(recycle_writer.AddRecord(Slice("bar")));
+  ASSERT_GE(get_reader_contents()->size(), log::kBlockSize * 2);
+  CheckRecordAndTimestampSize("foooo", ts_sz_two);
+  CheckRecordAndTimestampSize("bar", ts_sz_two);
+  ASSERT_EQ("EOF", Read());
+}
+
 // Do NOT enable compression for this instantiation.
 INSTANTIATE_TEST_CASE_P(
    Log, LogTest,
@@ -940,6 +1032,35 @@ TEST_P(CompressionLogTest, ReadWrite) {
  ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
 }

+TEST_P(CompressionLogTest, ReadWriteWithTimestampSize) {
+  CompressionType compression_type = std::get<2>(GetParam());
+  if (!StreamingCompressionTypeSupported(compression_type)) {
+    ROCKSDB_GTEST_SKIP("Test requires support for compression type");
+    return;
+  }
+  ASSERT_OK(SetupTestEnv());
+  std::unordered_map<uint32_t, size_t> ts_sz_one = {
+      {1, sizeof(uint64_t)},
+  };
+  Write("foo", &ts_sz_one);
+  Write("bar");
+  std::unordered_map<uint32_t, size_t> ts_sz_two = {{2, sizeof(char)}};
+  Write("", &ts_sz_two);
+  Write("xxxx");
+
+  CheckRecordAndTimestampSize("foo", ts_sz_one);
+  CheckRecordAndTimestampSize("bar", ts_sz_one);
+  std::unordered_map<uint32_t, size_t> expected_ts_sz_two;
+  // User-defined timestamp size records are accumulated and applied to
+  // subsequent records.
+  expected_ts_sz_two.insert(ts_sz_one.begin(), ts_sz_one.end());
+  expected_ts_sz_two.insert(ts_sz_two.begin(), ts_sz_two.end());
+  CheckRecordAndTimestampSize("", expected_ts_sz_two);
+  CheckRecordAndTimestampSize("xxxx", expected_ts_sz_two);
+  ASSERT_EQ("EOF", Read());
+  ASSERT_EQ("EOF", Read());  // Make sure reads at eof work
+}
+
 TEST_P(CompressionLogTest, ManyBlocks) {
  CompressionType compression_type = std::get<2>(GetParam());
  if (!StreamingCompressionTypeSupported(compression_type)) {
@@ -16,6 +16,7 @@
 #include "rocksdb/io_status.h"
 #include "util/coding.h"
 #include "util/crc32c.h"
+#include "util/udt_util.h"

 namespace ROCKSDB_NAMESPACE {
 namespace log {
@@ -73,7 +74,6 @@ IOStatus Writer::AddRecord(const Slice& slice,
  // Fragment the record if necessary and emit it.  Note that if slice
  // is empty, we still want to iterate once to emit a single
  // zero-length record
-  IOStatus s;
  bool begin = true;
  int compress_remaining = 0;
  bool compress_start = false;
@@ -81,6 +81,8 @@ IOStatus Writer::AddRecord(const Slice& slice,
    compress_->Reset();
    compress_start = true;
  }
+
+  IOStatus s;
  do {
    const int64_t leftover = kBlockSize - block_offset_;
    assert(leftover >= 0);
@@ -194,6 +196,33 @@ IOStatus Writer::AddCompressionTypeRecord() {
  return s;
 }

+IOStatus Writer::MaybeAddUserDefinedTimestampSizeRecord(
+    const std::unordered_map<uint32_t, size_t>& cf_to_ts_sz,
+    Env::IOPriority rate_limiter_priority) {
+  std::vector<std::pair<uint32_t, size_t>> ts_sz_to_record;
+  for (const auto& [cf_id, ts_sz] : cf_to_ts_sz) {
+    if (recorded_cf_to_ts_sz_.count(cf_id) != 0) {
+      // A column family's user-defined timestamp size should not be
+      // updated while DB is running.
+      assert(recorded_cf_to_ts_sz_[cf_id] == ts_sz);
+    } else if (ts_sz != 0) {
+      ts_sz_to_record.emplace_back(cf_id, ts_sz);
+      recorded_cf_to_ts_sz_.insert(std::make_pair(cf_id, ts_sz));
+    }
+  }
+  if (ts_sz_to_record.empty()) {
+    return IOStatus::OK();
+  }
+
+  UserDefinedTimestampSizeRecord record(std::move(ts_sz_to_record));
+  std::string encoded;
+  record.EncodeTo(&encoded);
+  RecordType type = recycle_log_files_ ? kRecyclableUserDefinedTimestampSizeType
+                                       : kUserDefinedTimestampSizeType;
+  return EmitPhysicalRecord(type, encoded.data(), encoded.size(),
+                            rate_limiter_priority);
+}
+
 bool Writer::BufferIsEmpty() { return dest_->BufferIsEmpty(); }

 IOStatus Writer::EmitPhysicalRecord(RecordType t, const char* ptr, size_t n,
@@ -209,7 +238,8 @@ IOStatus Writer::EmitPhysicalRecord(RecordType t, const char* ptr, size_t n,
  buf[6] = static_cast<char>(t);

  uint32_t crc = type_crc_[t];
-  if (t < kRecyclableFullType || t == kSetCompressionType) {
+  if (t < kRecyclableFullType || t == kSetCompressionType ||
+      t == kUserDefinedTimestampSizeType) {
    // Legacy record format
    assert(block_offset_ + kHeaderSize + n <= kBlockSize);
    header_size = kHeaderSize;
@@ -10,6 +10,8 @@

 #include <cstdint>
 #include <memory>
+#include <unordered_map>
+#include <vector>

 #include "db/log_format.h"
 #include "rocksdb/compression_type.h"
@@ -87,6 +89,15 @@ class Writer {
                     Env::IOPriority rate_limiter_priority = Env::IO_TOTAL);
  IOStatus AddCompressionTypeRecord();

+  // If there are column families in `cf_to_ts_sz` not included in
+  // `recorded_cf_to_ts_sz_` and its user-defined timestamp size is non-zero,
+  // adds a record of type kUserDefinedTimestampSizeType or
+  // kRecyclableUserDefinedTimestampSizeType for these column families.
+  // This timestamp size record applies to all subsequent records.
+  IOStatus MaybeAddUserDefinedTimestampSizeRecord(
+      const std::unordered_map<uint32_t, size_t>& cf_to_ts_sz,
+      Env::IOPriority rate_limiter_priority = Env::IO_TOTAL);
+
  WritableFileWriter* file() { return dest_.get(); }
  const WritableFileWriter* file() const { return dest_.get(); }

@@ -122,6 +133,11 @@ class Writer {
  StreamingCompress* compress_;
  // Reusable compressed output buffer
  std::unique_ptr<char[]> compressed_buffer_;
+
+  // The recorded user-defined timestamp size that have been written so far.
+  // Since the user-defined timestamp size cannot be changed while the DB is
+  // running, existing entry in this map cannot be updated.
+  std::unordered_map<uint32_t, size_t> recorded_cf_to_ts_sz_;
 };

 }  // namespace log
@@ -26,7 +26,7 @@
 #include "memory/arena.h"
 #include "memory/memory_usage.h"
 #include "monitoring/perf_context_imp.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "port/lang.h"
 #include "port/port.h"
 #include "rocksdb/comparator.h"
@@ -256,7 +256,7 @@ void MemTable::UpdateOldestKeyTime() {
 }

 Status MemTable::VerifyEntryChecksum(const char* entry,
-                                     size_t protection_bytes_per_key,
+                                     uint32_t protection_bytes_per_key,
                                     bool allow_data_in_errors) {
  if (protection_bytes_per_key == 0) {
    return Status::OK();
@@ -285,28 +285,11 @@ Status MemTable::VerifyEntryChecksum(const char* entry,
  Slice value = Slice(value_ptr, value_length);

  const char* checksum_ptr = value_ptr + value_length;
-  uint64_t expected = ProtectionInfo64()
-                          .ProtectKVO(user_key, value, type)
-                          .ProtectS(seq)
-                          .GetVal();
-  bool match = true;
-  switch (protection_bytes_per_key) {
-    case 1:
-      match = static_cast<uint8_t>(checksum_ptr[0]) ==
-              static_cast<uint8_t>(expected);
-      break;
-    case 2:
-      match = DecodeFixed16(checksum_ptr) == static_cast<uint16_t>(expected);
-      break;
-    case 4:
-      match = DecodeFixed32(checksum_ptr) == static_cast<uint32_t>(expected);
-      break;
-    case 8:
-      match = DecodeFixed64(checksum_ptr) == expected;
-      break;
-    default:
-      assert(false);
-  }
+  bool match =
+      ProtectionInfo64()
+          .ProtectKVO(user_key, value, type)
+          .ProtectS(seq)
+          .Verify(static_cast<uint8_t>(protection_bytes_per_key), checksum_ptr);
  if (!match) {
    std::string msg(
        "Corrupted memtable entry, per key-value checksum verification "
@@ -526,7 +509,7 @@ class MemTableIterator : public InternalIterator {
  bool valid_;
  bool arena_mode_;
  bool value_pinned_;
-  size_t protection_bytes_per_key_;
+  uint32_t protection_bytes_per_key_;
  Status status_;
  Logger* logger_;

@@ -684,28 +667,15 @@ void MemTable::UpdateEntryChecksum(const ProtectionInfoKVOS64* kv_prot_info,
    return;
  }

-  uint64_t checksum = 0;
  if (kv_prot_info == nullptr) {
-    checksum =
-        ProtectionInfo64().ProtectKVO(key, value, type).ProtectS(s).GetVal();
+    ProtectionInfo64()
+        .ProtectKVO(key, value, type)
+        .ProtectS(s)
+        .Encode(static_cast<uint8_t>(moptions_.protection_bytes_per_key),
+                checksum_ptr);
  } else {
-    checksum = kv_prot_info->GetVal();
-  }
-  switch (moptions_.protection_bytes_per_key) {
-    case 1:
-      checksum_ptr[0] = static_cast<uint8_t>(checksum);
-      break;
-    case 2:
-      EncodeFixed16(checksum_ptr, static_cast<uint16_t>(checksum));
-      break;
-    case 4:
-      EncodeFixed32(checksum_ptr, static_cast<uint32_t>(checksum));
-      break;
-    case 8:
-      EncodeFixed64(checksum_ptr, checksum);
-      break;
-    default:
-      assert(false);
+    kv_prot_info->Encode(
+        static_cast<uint8_t>(moptions_.protection_bytes_per_key), checksum_ptr);
  }
 }

@@ -902,7 +872,7 @@ struct Saver {
  ReadCallback* callback_;
  bool* is_blob_index;
  bool allow_data_in_errors;
-  size_t protection_bytes_per_key;
+  uint32_t protection_bytes_per_key;
  bool CheckCallback(SequenceNumber _seq) {
    if (callback_) {
      return callback_->IsVisible(_seq);
@@ -529,7 +529,7 @@ class MemTable {

  // Returns Corruption status if verification fails.
  static Status VerifyEntryChecksum(const char* entry,
-                                    size_t protection_bytes_per_key,
+                                    uint32_t protection_bytes_per_key,
                                    bool allow_data_in_errors = false);

 private:
@@ -15,7 +15,7 @@
 #include "db/wide/wide_column_serialization.h"
 #include "logging/logging.h"
 #include "monitoring/perf_context_imp.h"
-#include "monitoring/statistics.h"
+#include "monitoring/statistics_impl.h"
 #include "port/likely.h"
 #include "rocksdb/comparator.h"
 #include "rocksdb/db.h"
@@ -518,8 +518,9 @@ class Repairer {
    if (status.ok()) {
      // TODO: plumb Env::IOActivity
      const ReadOptions read_options;
-      status = table_cache_->GetTableProperties(file_options_, read_options,
-                                                icmp_, t->meta, &props);
+      status = table_cache_->GetTableProperties(
+          file_options_, read_options, icmp_, t->meta, &props,
+          0 /* block_protection_bytes_per_key */);
    }
    if (status.ok()) {
      auto s =
@@ -550,6 +551,17 @@ class Repairer {
      }
      t->meta.oldest_ancester_time = props->creation_time;
    }
+    if (status.ok()) {
+      uint64_t tail_size = 0;
+      bool contain_no_data_blocks =
+          props->num_entries > 0 &&
+          (props->num_entries == props->num_range_deletions);
+      if (props->tail_start_offset > 0 || contain_no_data_blocks) {
+        assert(props->tail_start_offset <= file_size);
+        tail_size = file_size - props->tail_start_offset;
+      }
+      t->meta.tail_size = tail_size;
+    }
    ColumnFamilyData* cfd = nullptr;
    if (status.ok()) {
      cfd = vset_.GetColumnFamilySet()->GetColumnFamily(t->column_family_id);
@@ -577,7 +589,8 @@ class Repairer {
          /*level=*/-1, /*max_file_size_for_l0_meta_pin=*/0,
          /*smallest_compaction_key=*/nullptr,
          /*largest_compaction_key=*/nullptr,
-          /*allow_unprepared_value=*/false);
+          /*allow_unprepared_value=*/false,
+          cfd->GetLatestMutableCFOptions()->block_protection_bytes_per_key);
      ParsedInternalKey parsed;
      for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
        Slice key = iter->key();
@@ -617,7 +630,9 @@ class Repairer {
      ReadOptions ropts;
      std::unique_ptr<FragmentedRangeTombstoneIterator> r_iter;
      status = table_cache_->GetRangeTombstoneIterator(
-          ropts, cfd->internal_comparator(), t->meta, &r_iter);
+          ropts, cfd->internal_comparator(), t->meta,
+          cfd->GetLatestMutableCFOptions()->block_protection_bytes_per_key,
+          &r_iter);

      if (r_iter) {
        r_iter->SeekToFirst();
@@ -678,7 +693,7 @@ class Repairer {
            table->meta.oldest_ancester_time, table->meta.file_creation_time,
            table->meta.epoch_number, table->meta.file_checksum,
            table->meta.file_checksum_func_name, table->meta.unique_id,
-            table->meta.compensated_range_deletion_size);
+            table->meta.compensated_range_deletion_size, table->meta.tail_size);
      }
      s = dummy_version_builder.Apply(&dummy_edit);
      if (s.ok()) {
@@ -93,7 +93,7 @@ TEST_F(SeqnoTimeTest, TemperatureBasicUniversal) {
    }
    ASSERT_OK(Flush());
  }
-  ASSERT_OK(dbfull()->WaitForCompact(true));
+  ASSERT_OK(dbfull()->WaitForCompact());

  // All data is hot, only output to penultimate level
  ASSERT_EQ("0,0,0,0,0,1", FilesPerLevel());
@@ -114,7 +114,7 @@ TEST_F(SeqnoTimeTest, TemperatureBasicUniversal) {
      });
    }
    ASSERT_OK(Flush());
-    ASSERT_OK(dbfull()->WaitForCompact(true));
+    ASSERT_OK(dbfull()->WaitForCompact());
    ASSERT_GT(GetSstSizeHelper(Temperature::kUnknown), 0);
    ASSERT_EQ(GetSstSizeHelper(Temperature::kCold), 0);
  }
@@ -128,7 +128,7 @@ TEST_F(SeqnoTimeTest, TemperatureBasicUniversal) {
      });
    }
    ASSERT_OK(Flush());
-    ASSERT_OK(dbfull()->WaitForCompact(true));
+    ASSERT_OK(dbfull()->WaitForCompact());
  }

  CompactRangeOptions cro;
@@ -91,7 +91,8 @@ Status TableCache::GetTableReader(
    const ReadOptions& ro, const FileOptions& file_options,
    const InternalKeyComparator& internal_comparator,
    const FileMetaData& file_meta, bool sequential_mode, bool record_read_stats,
-    HistogramImpl* file_read_hist, std::unique_ptr<TableReader>* table_reader,
+    uint8_t block_protection_bytes_per_key, HistogramImpl* file_read_hist,
+    std::unique_ptr<TableReader>* table_reader,
    const std::shared_ptr<const SliceTransform>& prefix_extractor,
    bool skip_filters, int level, bool prefetch_index_and_filter_in_cache,
    size_t max_file_size_for_l0_meta_pin, Temperature file_temperature) {
@@ -139,12 +140,13 @@ Status TableCache::GetTableReader(
    }
    s = ioptions_.table_factory->NewTableReader(
        ro,
-        TableReaderOptions(ioptions_, prefix_extractor, file_options,
-                           internal_comparator, skip_filters, immortal_tables_,
-                           false /* force_direct_prefetch */, level,
-                           block_cache_tracer_, max_file_size_for_l0_meta_pin,
-                           db_session_id_, file_meta.fd.GetNumber(),
-                           expected_unique_id, file_meta.fd.largest_seqno),
+        TableReaderOptions(
+            ioptions_, prefix_extractor, file_options, internal_comparator,
+            block_protection_bytes_per_key, skip_filters, immortal_tables_,
+            false /* force_direct_prefetch */, level, block_cache_tracer_,
+            max_file_size_for_l0_meta_pin, db_session_id_,
+            file_meta.fd.GetNumber(), expected_unique_id,
+            file_meta.fd.largest_seqno, file_meta.tail_size),
        std::move(file_reader), file_meta.fd.GetFileSize(), table_reader,
        prefetch_index_and_filter_in_cache);
    TEST_SYNC_POINT("TableCache::GetTableReader:0");
@@ -156,6 +158,7 @@ Status TableCache::FindTable(
    const ReadOptions& ro, const FileOptions& file_options,
    const InternalKeyComparator& internal_comparator,
    const FileMetaData& file_meta, TypedHandle** handle,
+    uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor,
    const bool no_io, bool record_read_stats, HistogramImpl* file_read_hist,
    bool skip_filters, int level, bool prefetch_index_and_filter_in_cache,
@@ -179,12 +182,12 @@ Status TableCache::FindTable(
    }

    std::unique_ptr<TableReader> table_reader;
-    Status s =
-        GetTableReader(ro, file_options, internal_comparator, file_meta,
-                       false /* sequential mode */, record_read_stats,
-                       file_read_hist, &table_reader, prefix_extractor,
-                       skip_filters, level, prefetch_index_and_filter_in_cache,
-                       max_file_size_for_l0_meta_pin, file_temperature);
+    Status s = GetTableReader(ro, file_options, internal_comparator, file_meta,
+                              false /* sequential mode */, record_read_stats,
+                              block_protection_bytes_per_key, file_read_hist,
+                              &table_reader, prefix_extractor, skip_filters,
+                              level, prefetch_index_and_filter_in_cache,
+                              max_file_size_for_l0_meta_pin, file_temperature);
    if (!s.ok()) {
      assert(table_reader == nullptr);
      RecordTick(ioptions_.stats, NO_FILE_ERRORS);
@@ -212,6 +215,7 @@ InternalIterator* TableCache::NewIterator(
    size_t max_file_size_for_l0_meta_pin,
    const InternalKey* smallest_compaction_key,
    const InternalKey* largest_compaction_key, bool allow_unprepared_value,
+    uint8_t block_protection_bytes_per_key,
    TruncatedRangeDelIterator** range_del_iter) {
  PERF_TIMER_GUARD(new_table_iterator_nanos);

@@ -225,12 +229,13 @@ InternalIterator* TableCache::NewIterator(
  auto& fd = file_meta.fd;
  table_reader = fd.table_reader;
  if (table_reader == nullptr) {
-    s = FindTable(
-        options, file_options, icomparator, file_meta, &handle,
-        prefix_extractor, options.read_tier == kBlockCacheTier /* no_io */,
-        !for_compaction /* record_read_stats */, file_read_hist, skip_filters,
-        level, true /* prefetch_index_and_filter_in_cache */,
-        max_file_size_for_l0_meta_pin, file_meta.temperature);
+    s = FindTable(options, file_options, icomparator, file_meta, &handle,
+                  block_protection_bytes_per_key, prefix_extractor,
+                  options.read_tier == kBlockCacheTier /* no_io */,
+                  !for_compaction /* record_read_stats */, file_read_hist,
+                  skip_filters, level,
+                  true /* prefetch_index_and_filter_in_cache */,
+                  max_file_size_for_l0_meta_pin, file_meta.temperature);
    if (s.ok()) {
      table_reader = cache_.Value(handle);
    }
@@ -308,7 +313,7 @@ InternalIterator* TableCache::NewIterator(
 Status TableCache::GetRangeTombstoneIterator(
    const ReadOptions& options,
    const InternalKeyComparator& internal_comparator,
-    const FileMetaData& file_meta,
+    const FileMetaData& file_meta, uint8_t block_protection_bytes_per_key,
    std::unique_ptr<FragmentedRangeTombstoneIterator>* out_iter) {
  assert(out_iter);
  const FileDescriptor& fd = file_meta.fd;
@@ -317,7 +322,7 @@ Status TableCache::GetRangeTombstoneIterator(
  TypedHandle* handle = nullptr;
  if (t == nullptr) {
    s = FindTable(options, file_options_, internal_comparator, file_meta,
-                  &handle);
+                  &handle, block_protection_bytes_per_key);
    if (s.ok()) {
      t = cache_.Value(handle);
    }
@@ -403,6 +408,7 @@ Status TableCache::Get(
    const ReadOptions& options,
    const InternalKeyComparator& internal_comparator,
    const FileMetaData& file_meta, const Slice& k, GetContext* get_context,
+    uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor,
    HistogramImpl* file_read_hist, bool skip_filters, int level,
    size_t max_file_size_for_l0_meta_pin) {
@@ -430,7 +436,7 @@ Status TableCache::Get(
    assert(s.ok());
    if (t == nullptr) {
      s = FindTable(options, file_options_, internal_comparator, file_meta,
-                    &handle, prefix_extractor,
+                    &handle, block_protection_bytes_per_key, prefix_extractor,
                    options.read_tier == kBlockCacheTier /* no_io */,
                    true /* record_read_stats */, file_read_hist, skip_filters,
                    level, true /* prefetch_index_and_filter_in_cache */,
@@ -513,7 +519,8 @@ Status TableCache::MultiGetFilter(
    const FileMetaData& file_meta,
    const std::shared_ptr<const SliceTransform>& prefix_extractor,
    HistogramImpl* file_read_hist, int level,
-    MultiGetContext::Range* mget_range, TypedHandle** table_handle) {
+    MultiGetContext::Range* mget_range, TypedHandle** table_handle,
+    uint8_t block_protection_bytes_per_key) {
  auto& fd = file_meta.fd;
  IterKey row_cache_key;
  std::string row_cache_entry_buffer;
@@ -531,12 +538,13 @@ Status TableCache::MultiGetFilter(
  MultiGetContext::Range tombstone_range(*mget_range, mget_range->begin(),
                                         mget_range->end());
  if (t == nullptr) {
-    s = FindTable(
-        options, file_options_, internal_comparator, file_meta, &handle,
-        prefix_extractor, options.read_tier == kBlockCacheTier /* no_io */,
-        true /* record_read_stats */, file_read_hist, /*skip_filters=*/false,
-        level, true /* prefetch_index_and_filter_in_cache */,
-        /*max_file_size_for_l0_meta_pin=*/0, file_meta.temperature);
+    s = FindTable(options, file_options_, internal_comparator, file_meta,
+                  &handle, block_protection_bytes_per_key, prefix_extractor,
+                  options.read_tier == kBlockCacheTier /* no_io */,
+                  true /* record_read_stats */, file_read_hist,
+                  /*skip_filters=*/false, level,
+                  true /* prefetch_index_and_filter_in_cache */,
+                  /*max_file_size_for_l0_meta_pin=*/0, file_meta.temperature);
    if (s.ok()) {
      t = cache_.Value(handle);
    }
@@ -564,6 +572,7 @@ Status TableCache::GetTableProperties(
    const InternalKeyComparator& internal_comparator,
    const FileMetaData& file_meta,
    std::shared_ptr<const TableProperties>* properties,
+    uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor, bool no_io) {
  auto table_reader = file_meta.fd.table_reader;
  // table already been pre-loaded?
@@ -575,7 +584,8 @@ Status TableCache::GetTableProperties(

  TypedHandle* table_handle = nullptr;
  Status s = FindTable(read_options, file_options, internal_comparator,
-                       file_meta, &table_handle, prefix_extractor, no_io);
+                       file_meta, &table_handle, block_protection_bytes_per_key,
+                       prefix_extractor, no_io);
  if (!s.ok()) {
    return s;
  }
@@ -588,12 +598,14 @@ Status TableCache::GetTableProperties(

 Status TableCache::ApproximateKeyAnchors(
    const ReadOptions& ro, const InternalKeyComparator& internal_comparator,
-    const FileMetaData& file_meta, std::vector<TableReader::Anchor>& anchors) {
+    const FileMetaData& file_meta, uint8_t block_protection_bytes_per_key,
+    std::vector<TableReader::Anchor>& anchors) {
  Status s;
  TableReader* t = file_meta.fd.table_reader;
  TypedHandle* handle = nullptr;
  if (t == nullptr) {
-    s = FindTable(ro, file_options_, internal_comparator, file_meta, &handle);
+    s = FindTable(ro, file_options_, internal_comparator, file_meta, &handle,
+                  block_protection_bytes_per_key);
    if (s.ok()) {
      t = cache_.Value(handle);
    }
@@ -610,7 +622,7 @@ Status TableCache::ApproximateKeyAnchors(
 size_t TableCache::GetMemoryUsageByTableReader(
    const FileOptions& file_options, const ReadOptions& read_options,
    const InternalKeyComparator& internal_comparator,
-    const FileMetaData& file_meta,
+    const FileMetaData& file_meta, uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor) {
  auto table_reader = file_meta.fd.table_reader;
  // table already been pre-loaded?
@@ -620,7 +632,8 @@ size_t TableCache::GetMemoryUsageByTableReader(

  TypedHandle* table_handle = nullptr;
  Status s = FindTable(read_options, file_options, internal_comparator,
-                       file_meta, &table_handle, prefix_extractor, true);
+                       file_meta, &table_handle, block_protection_bytes_per_key,
+                       prefix_extractor, true /* no_io */);
  if (!s.ok()) {
    return 0;
  }
@@ -639,16 +652,17 @@ uint64_t TableCache::ApproximateOffsetOf(
    const ReadOptions& read_options, const Slice& key,
    const FileMetaData& file_meta, TableReaderCaller caller,
    const InternalKeyComparator& internal_comparator,
+    uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor) {
  uint64_t result = 0;
  TableReader* table_reader = file_meta.fd.table_reader;
  TypedHandle* table_handle = nullptr;
  if (table_reader == nullptr) {
    const bool for_compaction = (caller == TableReaderCaller::kCompaction);
-    Status s =
-        FindTable(read_options, file_options_, internal_comparator, file_meta,
-                  &table_handle, prefix_extractor, false /* no_io */,
-                  !for_compaction /* record_read_stats */);
+    Status s = FindTable(
+        read_options, file_options_, internal_comparator, file_meta,
+        &table_handle, block_protection_bytes_per_key, prefix_extractor,
+        false /* no_io */, !for_compaction /* record_read_stats */);
    if (s.ok()) {
      table_reader = cache_.Value(table_handle);
    }
@@ -668,16 +682,17 @@ uint64_t TableCache::ApproximateSize(
    const ReadOptions& read_options, const Slice& start, const Slice& end,
    const FileMetaData& file_meta, TableReaderCaller caller,
    const InternalKeyComparator& internal_comparator,
+    uint8_t block_protection_bytes_per_key,
    const std::shared_ptr<const SliceTransform>& prefix_extractor) {
  uint64_t result = 0;
  TableReader* table_reader = file_meta.fd.table_reader;
  TypedHandle* table_handle = nullptr;
  if (table_reader == nullptr) {
    const bool for_compaction = (caller == TableReaderCaller::kCompaction);
-    Status s =
-        FindTable(read_options, file_options_, internal_comparator, file_meta,
-                  &table_handle, prefix_extractor, false /* no_io */,
-                  !for_compaction /* record_read_stats */);
+    Status s = FindTable(
+        read_options, file_options_, internal_comparator, file_meta,
+        &table_handle, block_protection_bytes_per_key, prefix_extractor,
+        false /* no_io */, !for_compaction /* record_read_stats */);
    if (s.ok()) {
      table_reader = cache_.Value(table_handle);
    }
@@ -96,6 +96,7 @@ class TableCache {
      size_t max_file_size_for_l0_meta_pin,
      const InternalKey* smallest_compaction_key,
      const InternalKey* largest_compaction_key, bool allow_unprepared_value,
+      uint8_t protection_bytes_per_key,
      TruncatedRangeDelIterator** range_del_iter = nullptr);

  // If a seek to internal key "k" in specified file finds an entry,
@@ -112,6 +113,7 @@ class TableCache {
      const ReadOptions& options,
      const InternalKeyComparator& internal_comparator,
      const FileMetaData& file_meta, const Slice& k, GetContext* get_context,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr,
      HistogramImpl* file_read_hist = nullptr, bool skip_filters = false,
      int level = -1, size_t max_file_size_for_l0_meta_pin = 0);
@@ -121,7 +123,7 @@ class TableCache {
  Status GetRangeTombstoneIterator(
      const ReadOptions& options,
      const InternalKeyComparator& internal_comparator,
-      const FileMetaData& file_meta,
+      const FileMetaData& file_meta, uint8_t block_protection_bytes_per_key,
      std::unique_ptr<FragmentedRangeTombstoneIterator>* out_iter);

  // Call table reader's MultiGetFilter to use the bloom filter to filter out
@@ -135,7 +137,8 @@ class TableCache {
      const FileMetaData& file_meta,
      const std::shared_ptr<const SliceTransform>& prefix_extractor,
      HistogramImpl* file_read_hist, int level,
-      MultiGetContext::Range* mget_range, TypedHandle** table_handle);
+      MultiGetContext::Range* mget_range, TypedHandle** table_handle,
+      uint8_t block_protection_bytes_per_key);

  // If a seek to internal key "k" in specified file finds an entry,
  // call get_context->SaveValue() repeatedly until
@@ -150,6 +153,7 @@ class TableCache {
      Status, MultiGet, const ReadOptions& options,
      const InternalKeyComparator& internal_comparator,
      const FileMetaData& file_meta, const MultiGetContext::Range* mget_range,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr,
      HistogramImpl* file_read_hist = nullptr, bool skip_filters = false,
      bool skip_range_deletions = false, int level = -1,
@@ -165,6 +169,7 @@ class TableCache {
      const ReadOptions& ro, const FileOptions& toptions,
      const InternalKeyComparator& internal_comparator,
      const FileMetaData& file_meta, TypedHandle**,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr,
      const bool no_io = false, bool record_read_stats = true,
      HistogramImpl* file_read_hist = nullptr, bool skip_filters = false,
@@ -183,12 +188,14 @@ class TableCache {
      const InternalKeyComparator& internal_comparator,
      const FileMetaData& file_meta,
      std::shared_ptr<const TableProperties>* properties,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr,
      bool no_io = false);

  Status ApproximateKeyAnchors(const ReadOptions& ro,
                               const InternalKeyComparator& internal_comparator,
                               const FileMetaData& file_meta,
+                               uint8_t block_protection_bytes_per_key,
                               std::vector<TableReader::Anchor>& anchors);

  // Return total memory usage of the table reader of the file.
@@ -196,7 +203,7 @@ class TableCache {
  size_t GetMemoryUsageByTableReader(
      const FileOptions& toptions, const ReadOptions& read_options,
      const InternalKeyComparator& internal_comparator,
-      const FileMetaData& file_meta,
+      const FileMetaData& file_meta, uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr);

  // Returns approximated offset of a key in a file represented by fd.
@@ -204,6 +211,7 @@ class TableCache {
      const ReadOptions& read_options, const Slice& key,
      const FileMetaData& file_meta, TableReaderCaller caller,
      const InternalKeyComparator& internal_comparator,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr);

  // Returns approximated data size between start and end keys in a file
@@ -212,6 +220,7 @@ class TableCache {
      const ReadOptions& read_options, const Slice& start, const Slice& end,
      const FileMetaData& file_meta, TableReaderCaller caller,
      const InternalKeyComparator& internal_comparator,
+      uint8_t block_protection_bytes_per_key,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr);

  CacheInterface& get_cache() { return cache_; }
@@ -234,8 +243,8 @@ class TableCache {
      const ReadOptions& ro, const FileOptions& file_options,
      const InternalKeyComparator& internal_comparator,
      const FileMetaData& file_meta, bool sequential_mode,
-      bool record_read_stats, HistogramImpl* file_read_hist,
-      std::unique_ptr<TableReader>* table_reader,
+      bool record_read_stats, uint8_t block_protection_bytes_per_key,
+      HistogramImpl* file_read_hist, std::unique_ptr<TableReader>* table_reader,
      const std::shared_ptr<const SliceTransform>& prefix_extractor = nullptr,
      bool skip_filters = false, int level = -1,
      bool prefetch_index_and_filter_in_cache = true,
@@ -17,6 +17,7 @@ namespace ROCKSDB_NAMESPACE {
 DEFINE_SYNC_AND_ASYNC(Status, TableCache::MultiGet)
 (const ReadOptions& options, const InternalKeyComparator& internal_comparator,
 const FileMetaData& file_meta, const MultiGetContext::Range* mget_range,
+ uint8_t block_protection_bytes_per_key,
 const std::shared_ptr<const SliceTransform>& prefix_extractor,
 HistogramImpl* file_read_hist, bool skip_filters, bool skip_range_deletions,
 int level, TypedHandle* handle) {
@@ -65,7 +66,7 @@ DEFINE_SYNC_AND_ASYNC(Status, TableCache::MultiGet)
    if (t == nullptr) {
      assert(handle == nullptr);
      s = FindTable(options, file_options_, internal_comparator, file_meta,
-                    &handle, prefix_extractor,
+                    &handle, block_protection_bytes_per_key, prefix_extractor,
                    options.read_tier == kBlockCacheTier /* no_io */,
                    true /* record_read_stats */, file_read_hist, skip_filters,
                    level, true /* prefetch_index_and_filter_in_cache */,
@@ -1257,7 +1257,8 @@ class VersionBuilder::Rep {
      InternalStats* internal_stats, int max_threads,
      bool prefetch_index_and_filter_in_cache, bool is_initial_load,
      const std::shared_ptr<const SliceTransform>& prefix_extractor,
-      size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options) {
+      size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options,
+      uint8_t block_protection_bytes_per_key) {
    assert(table_cache_ != nullptr);

    size_t table_cache_capacity =
@@ -1326,7 +1327,8 @@ class VersionBuilder::Rep {
        statuses[file_idx] = table_cache_->FindTable(
            read_options, file_options_,
            *(base_vstorage_->InternalComparator()), *file_meta, &handle,
-            prefix_extractor, false /*no_io */, true /* record_read_stats */,
+            block_protection_bytes_per_key, prefix_extractor, false /*no_io */,
+            true /* record_read_stats */,
            internal_stats->GetFileReadHist(level), false, level,
            prefetch_index_and_filter_in_cache, max_file_size_for_l0_meta_pin,
            file_meta->temperature);
@@ -1384,11 +1386,12 @@ Status VersionBuilder::LoadTableHandlers(
    InternalStats* internal_stats, int max_threads,
    bool prefetch_index_and_filter_in_cache, bool is_initial_load,
    const std::shared_ptr<const SliceTransform>& prefix_extractor,
-    size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options) {
-  return rep_->LoadTableHandlers(internal_stats, max_threads,
-                                 prefetch_index_and_filter_in_cache,
-                                 is_initial_load, prefix_extractor,
-                                 max_file_size_for_l0_meta_pin, read_options);
+    size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options,
+    uint8_t block_protection_bytes_per_key) {
+  return rep_->LoadTableHandlers(
+      internal_stats, max_threads, prefetch_index_and_filter_in_cache,
+      is_initial_load, prefix_extractor, max_file_size_for_l0_meta_pin,
+      read_options, block_protection_bytes_per_key);
 }

 uint64_t VersionBuilder::GetMinOldestBlobFileNumber() const {
@@ -48,7 +48,8 @@ class VersionBuilder {
      InternalStats* internal_stats, int max_threads,
      bool prefetch_index_and_filter_in_cache, bool is_initial_load,
      const std::shared_ptr<const SliceTransform>& prefix_extractor,
-      size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options);
+      size_t max_file_size_for_l0_meta_pin, const ReadOptions& read_options,
+      uint8_t block_protection_bytes_per_key);
  uint64_t GetMinOldestBlobFileNumber() const;

 private:
@@ -73,7 +73,7 @@ class VersionBuilderTest : public testing::Test {
        /* marked_for_compact */ false, Temperature::kUnknown,
        oldest_blob_file_number, kUnknownOldestAncesterTime,
        kUnknownFileCreationTime, epoch_number, kUnknownFileChecksum,
-        kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+        kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
    f->compensated_file_size = file_size;
    f->num_entries = num_entries;
    f->num_deletions = num_deletions;
@@ -136,7 +136,7 @@ class VersionBuilderTest : public testing::Test {
                  Temperature::kUnknown, blob_file_number,
                  kUnknownOldestAncesterTime, kUnknownFileCreationTime,
                  epoch_number, kUnknownFileChecksum,
-                  kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+                  kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  }

  void UpdateVersionStorageInfo(VersionStorageInfo* vstorage) {
@@ -183,11 +183,12 @@ TEST_F(VersionBuilderTest, ApplyAndSaveTo) {
  UpdateVersionStorageInfo();

  VersionEdit version_edit;
-  version_edit.AddFile(
-      2, 666, 0, 100U, GetInternalKey("301"), GetInternalKey("350"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(2, 666, 0, 100U, GetInternalKey("301"),
+                       GetInternalKey("350"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  version_edit.DeleteFile(3, 27U);

  EnvOptions env_options;
@@ -230,11 +231,12 @@ TEST_F(VersionBuilderTest, ApplyAndSaveToDynamic) {
  UpdateVersionStorageInfo();

  VersionEdit version_edit;
-  version_edit.AddFile(
-      3, 666, 0, 100U, GetInternalKey("301"), GetInternalKey("350"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(3, 666, 0, 100U, GetInternalKey("301"),
+                       GetInternalKey("350"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  version_edit.DeleteFile(0, 1U);
  version_edit.DeleteFile(0, 88U);
@@ -281,11 +283,12 @@ TEST_F(VersionBuilderTest, ApplyAndSaveToDynamic2) {
  UpdateVersionStorageInfo();

  VersionEdit version_edit;
-  version_edit.AddFile(
-      4, 666, 0, 100U, GetInternalKey("301"), GetInternalKey("350"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(4, 666, 0, 100U, GetInternalKey("301"),
+                       GetInternalKey("350"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  version_edit.DeleteFile(0, 1U);
  version_edit.DeleteFile(0, 88U);
  version_edit.DeleteFile(4, 6U);
@@ -317,31 +320,36 @@ TEST_F(VersionBuilderTest, ApplyMultipleAndSaveTo) {
  UpdateVersionStorageInfo();

  VersionEdit version_edit;
-  version_edit.AddFile(
-      2, 666, 0, 100U, GetInternalKey("301"), GetInternalKey("350"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 676, 0, 100U, GetInternalKey("401"), GetInternalKey("450"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 636, 0, 100U, GetInternalKey("601"), GetInternalKey("650"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 616, 0, 100U, GetInternalKey("501"), GetInternalKey("550"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 606, 0, 100U, GetInternalKey("701"), GetInternalKey("750"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(2, 666, 0, 100U, GetInternalKey("301"),
+                       GetInternalKey("350"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 676, 0, 100U, GetInternalKey("401"),
+                       GetInternalKey("450"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 636, 0, 100U, GetInternalKey("601"),
+                       GetInternalKey("650"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 616, 0, 100U, GetInternalKey("501"),
+                       GetInternalKey("550"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 606, 0, 100U, GetInternalKey("701"),
+                       GetInternalKey("750"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  EnvOptions env_options;
  constexpr TableCache* table_cache = nullptr;
@@ -376,46 +384,53 @@ TEST_F(VersionBuilderTest, ApplyDeleteAndSaveTo) {
                                  kCompactionStyleLevel, nullptr, false);

  VersionEdit version_edit;
-  version_edit.AddFile(
-      2, 666, 0, 100U, GetInternalKey("301"), GetInternalKey("350"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 676, 0, 100U, GetInternalKey("401"), GetInternalKey("450"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 636, 0, 100U, GetInternalKey("601"), GetInternalKey("650"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 616, 0, 100U, GetInternalKey("501"), GetInternalKey("550"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
-  version_edit.AddFile(
-      2, 606, 0, 100U, GetInternalKey("701"), GetInternalKey("750"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(2, 666, 0, 100U, GetInternalKey("301"),
+                       GetInternalKey("350"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 676, 0, 100U, GetInternalKey("401"),
+                       GetInternalKey("450"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 636, 0, 100U, GetInternalKey("601"),
+                       GetInternalKey("650"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 616, 0, 100U, GetInternalKey("501"),
+                       GetInternalKey("550"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
+  version_edit.AddFile(2, 606, 0, 100U, GetInternalKey("701"),
+                       GetInternalKey("750"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  ASSERT_OK(version_builder.Apply(&version_edit));

  VersionEdit version_edit2;
-  version_edit.AddFile(
-      2, 808, 0, 100U, GetInternalKey("901"), GetInternalKey("950"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(2, 808, 0, 100U, GetInternalKey("901"),
+                       GetInternalKey("950"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  version_edit2.DeleteFile(2, 616);
  version_edit2.DeleteFile(2, 636);
-  version_edit.AddFile(
-      2, 806, 0, 100U, GetInternalKey("801"), GetInternalKey("850"), 200, 200,
-      false, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  version_edit.AddFile(2, 806, 0, 100U, GetInternalKey("801"),
+                       GetInternalKey("850"), 200, 200, false,
+                       Temperature::kUnknown, kInvalidBlobFileNumber,
+                       kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                       kUnknownEpochNumber, kUnknownFileChecksum,
+                       kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  ASSERT_OK(version_builder.Apply(&version_edit2));
  ASSERT_OK(version_builder.SaveTo(&new_vstorage));
@@ -520,13 +535,14 @@ TEST_F(VersionBuilderTest, ApplyFileDeletionAndAddition) {

  constexpr bool marked_for_compaction = false;

-  addition.AddFile(
-      level, file_number, path_id, file_size,
-      GetInternalKey(smallest, smallest_seq),
-      GetInternalKey(largest, largest_seq), smallest_seqno, largest_seqno,
-      marked_for_compaction, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  addition.AddFile(level, file_number, path_id, file_size,
+                   GetInternalKey(smallest, smallest_seq),
+                   GetInternalKey(largest, largest_seq), smallest_seqno,
+                   largest_seqno, marked_for_compaction, Temperature::kUnknown,
+                   kInvalidBlobFileNumber, kUnknownOldestAncesterTime,
+                   kUnknownFileCreationTime, kUnknownEpochNumber,
+                   kUnknownFileChecksum, kUnknownFileChecksumFuncName,
+                   kNullUniqueId64x2, 0, 0);

  ASSERT_OK(builder.Apply(&addition));

@@ -570,12 +586,13 @@ TEST_F(VersionBuilderTest, ApplyFileAdditionAlreadyInBase) {
  constexpr SequenceNumber largest_seqno = 1000;
  constexpr bool marked_for_compaction = false;

-  edit.AddFile(
-      new_level, file_number, path_id, file_size, GetInternalKey(smallest),
-      GetInternalKey(largest), smallest_seqno, largest_seqno,
-      marked_for_compaction, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  edit.AddFile(new_level, file_number, path_id, file_size,
+               GetInternalKey(smallest), GetInternalKey(largest),
+               smallest_seqno, largest_seqno, marked_for_compaction,
+               Temperature::kUnknown, kInvalidBlobFileNumber,
+               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+               kUnknownEpochNumber, kUnknownFileChecksum,
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  const Status s = builder.Apply(&edit);
  ASSERT_TRUE(s.IsCorruption());
@@ -606,12 +623,13 @@ TEST_F(VersionBuilderTest, ApplyFileAdditionAlreadyApplied) {
  constexpr SequenceNumber largest_seqno = 1000;
  constexpr bool marked_for_compaction = false;

-  edit.AddFile(
-      level, file_number, path_id, file_size, GetInternalKey(smallest),
-      GetInternalKey(largest), smallest_seqno, largest_seqno,
-      marked_for_compaction, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  edit.AddFile(level, file_number, path_id, file_size, GetInternalKey(smallest),
+               GetInternalKey(largest), smallest_seqno, largest_seqno,
+               marked_for_compaction, Temperature::kUnknown,
+               kInvalidBlobFileNumber, kUnknownOldestAncesterTime,
+               kUnknownFileCreationTime, kUnknownEpochNumber,
+               kUnknownFileChecksum, kUnknownFileChecksumFuncName,
+               kNullUniqueId64x2, 0, 0);

  ASSERT_OK(builder.Apply(&edit));

@@ -619,12 +637,13 @@ TEST_F(VersionBuilderTest, ApplyFileAdditionAlreadyApplied) {

  constexpr int new_level = 2;

-  other_edit.AddFile(
-      new_level, file_number, path_id, file_size, GetInternalKey(smallest),
-      GetInternalKey(largest), smallest_seqno, largest_seqno,
-      marked_for_compaction, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  other_edit.AddFile(new_level, file_number, path_id, file_size,
+                     GetInternalKey(smallest), GetInternalKey(largest),
+                     smallest_seqno, largest_seqno, marked_for_compaction,
+                     Temperature::kUnknown, kInvalidBlobFileNumber,
+                     kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                     kUnknownEpochNumber, kUnknownFileChecksum,
+                     kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  const Status s = builder.Apply(&other_edit);
  ASSERT_TRUE(s.IsCorruption());
@@ -655,12 +674,13 @@ TEST_F(VersionBuilderTest, ApplyFileAdditionAndDeletion) {

  VersionEdit addition;

-  addition.AddFile(
-      level, file_number, path_id, file_size, GetInternalKey(smallest),
-      GetInternalKey(largest), smallest_seqno, largest_seqno,
-      marked_for_compaction, Temperature::kUnknown, kInvalidBlobFileNumber,
-      kUnknownOldestAncesterTime, kUnknownFileCreationTime, kUnknownEpochNumber,
-      kUnknownFileChecksum, kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+  addition.AddFile(level, file_number, path_id, file_size,
+                   GetInternalKey(smallest), GetInternalKey(largest),
+                   smallest_seqno, largest_seqno, marked_for_compaction,
+                   Temperature::kUnknown, kInvalidBlobFileNumber,
+                   kUnknownOldestAncesterTime, kUnknownFileCreationTime,
+                   kUnknownEpochNumber, kUnknownFileChecksum,
+                   kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  ASSERT_OK(builder.Apply(&addition));

@@ -1233,7 +1253,7 @@ TEST_F(VersionBuilderTest, SaveBlobFilesToConcurrentJobs) {
      GetInternalKey(largest), smallest_seqno, largest_seqno,
      marked_for_compaction, Temperature::kUnknown, blob_file_number,
      kUnknownOldestAncesterTime, kUnknownFileCreationTime, 2 /*epoch_number*/,
-      checksum_value, checksum_method, kNullUniqueId64x2, 0);
+      checksum_value, checksum_method, kNullUniqueId64x2, 0, 0);
  edit.AddBlobFile(blob_file_number, total_blob_count, total_blob_bytes,
                   checksum_method, checksum_value);

@@ -1321,7 +1341,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForBlobFiles) {
               /* oldest_blob_file_number */ 16, kUnknownOldestAncesterTime,
               kUnknownFileCreationTime, kUnknownEpochNumber,
               kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-               kNullUniqueId64x2, 0);
+               kNullUniqueId64x2, 0, 0);

  edit.AddFile(/* level */ 1, /* file_number */ 700, /* path_id */ 0,
               /* file_size */ 100, /* smallest */ GetInternalKey("801"),
@@ -1331,7 +1351,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForBlobFiles) {
               /* oldest_blob_file_number */ 1000, kUnknownOldestAncesterTime,
               kUnknownFileCreationTime, kUnknownEpochNumber,
               kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-               kNullUniqueId64x2, 0);
+               kNullUniqueId64x2, 0, 0);
  edit.AddBlobFile(/* blob_file_number */ 1000, /* total_blob_count */ 2000,
                   /* total_blob_bytes */ 200000,
                   /* checksum_method */ std::string(),
@@ -1552,7 +1572,7 @@ TEST_F(VersionBuilderTest, MaintainLinkedSstsForBlobFiles) {
      Temperature::kUnknown,
      /* oldest_blob_file_number */ 1, kUnknownOldestAncesterTime,
      kUnknownFileCreationTime, kUnknownEpochNumber, kUnknownFileChecksum,
-      kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+      kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  // Add an SST that does not reference any blob files.
  edit.AddFile(
@@ -1562,7 +1582,7 @@ TEST_F(VersionBuilderTest, MaintainLinkedSstsForBlobFiles) {
      /* largest_seqno */ 2200, /* marked_for_compaction */ false,
      Temperature::kUnknown, kInvalidBlobFileNumber, kUnknownOldestAncesterTime,
      kUnknownFileCreationTime, kUnknownEpochNumber, kUnknownFileChecksum,
-      kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+      kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  // Delete a file that references a blob file.
  edit.DeleteFile(/* level */ 1, /* file_number */ 6);
@@ -1585,7 +1605,7 @@ TEST_F(VersionBuilderTest, MaintainLinkedSstsForBlobFiles) {
               /* oldest_blob_file_number */ 3, kUnknownOldestAncesterTime,
               kUnknownFileCreationTime, kUnknownEpochNumber,
               kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-               kNullUniqueId64x2, 0);
+               kNullUniqueId64x2, 0, 0);

  // Trivially move a file that does not reference any blob files.
  edit.DeleteFile(/* level */ 1, /* file_number */ 13);
@@ -1597,7 +1617,7 @@ TEST_F(VersionBuilderTest, MaintainLinkedSstsForBlobFiles) {
               Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               kUnknownEpochNumber, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  // Add one more SST file that references a blob file, then promptly
  // delete it in a second version edit before the new version gets saved.
@@ -1611,7 +1631,7 @@ TEST_F(VersionBuilderTest, MaintainLinkedSstsForBlobFiles) {
               /* oldest_blob_file_number */ 5, kUnknownOldestAncesterTime,
               kUnknownFileCreationTime, kUnknownEpochNumber,
               kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-               kNullUniqueId64x2, 0);
+               kNullUniqueId64x2, 0, 0);

  VersionEdit edit2;

@@ -1712,7 +1732,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForL0FilesSortedByEpochNumber) {
      /* oldest_blob_file_number */ kInvalidBlobFileNumber,
      kUnknownOldestAncesterTime, kUnknownFileCreationTime,
      1 /* epoch_number */, kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-      kNullUniqueId64x2, 0);
+      kNullUniqueId64x2, 0, 0);
  version_edit_1.AddFile(
      /* level */ 0, /* file_number */ 2U, /* path_id */ 0,
      /* file_size */ 100, /* smallest */ GetInternalKey("b", 2),
@@ -1722,7 +1742,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForL0FilesSortedByEpochNumber) {
      /* oldest_blob_file_number */ kInvalidBlobFileNumber,
      kUnknownOldestAncesterTime, kUnknownFileCreationTime,
      1 /* epoch_number */, kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-      kNullUniqueId64x2, 0);
+      kNullUniqueId64x2, 0, 0);

  VersionBuilder version_builder_1(EnvOptions(), &ioptions_,
                                   nullptr /* table_cache */, &vstorage_,
@@ -1749,7 +1769,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForL0FilesSortedByEpochNumber) {
      /* oldest_blob_file_number */ kInvalidBlobFileNumber,
      kUnknownOldestAncesterTime, kUnknownFileCreationTime,
      1 /* epoch_number */, kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-      kNullUniqueId64x2, 0);
+      kNullUniqueId64x2, 0, 0);
  version_edit_2.AddFile(
      /* level */ 0, /* file_number */ 2U, /* path_id */ 0,
      /* file_size */ 100, /* smallest */ GetInternalKey("b", 2),
@@ -1759,7 +1779,7 @@ TEST_F(VersionBuilderTest, CheckConsistencyForL0FilesSortedByEpochNumber) {
      /* oldest_blob_file_number */ kInvalidBlobFileNumber,
      kUnknownOldestAncesterTime, kUnknownFileCreationTime,
      2 /* epoch_number */, kUnknownFileChecksum, kUnknownFileChecksumFuncName,
-      kNullUniqueId64x2, 0);
+      kNullUniqueId64x2, 0, 0);

  VersionBuilder version_builder_2(EnvOptions(), &ioptions_,
                                   nullptr /* table_cache */, &vstorage_,
@@ -238,6 +238,12 @@ bool VersionEdit::EncodeTo(std::string* dst) const {
                  f.compensated_range_deletion_size);
      PutLengthPrefixedSlice(dst, Slice(compensated_range_deletion_size));
    }
+    if (f.tail_size) {
+      PutVarint32(dst, NewFileCustomTag::kTailSize);
+      std::string varint_tail_size;
+      PutVarint64(&varint_tail_size, f.tail_size);
+      PutLengthPrefixedSlice(dst, Slice(varint_tail_size));
+    }

    TEST_SYNC_POINT_CALLBACK("VersionEdit::EncodeTo:NewFile4:CustomizeFields",
                             dst);
@@ -416,6 +422,11 @@ const char* VersionEdit::DecodeNewFile4From(Slice* input) {
            return "Invalid compensated range deletion size";
          }
          break;
+        case kTailSize:
+          if (!GetVarint64(&field, &f.tail_size)) {
+            return "invalid tail start offset";
+          }
+          break;
        default:
          if ((custom_tag & kCustomTagNonSafeIgnoreMask) != 0) {
            // Should not proceed if cannot understand it
@@ -851,6 +862,8 @@ std::string VersionEdit::DebugString(bool hex_key) const {
      InternalUniqueIdToExternal(&id);
      r.append(UniqueIdToHumanString(EncodeUniqueIdBytes(&id)));
    }
+    r.append(" tail size:");
+    AppendNumberTo(&r, f.tail_size);
  }

  for (const auto& blob_file_addition : blob_file_additions_) {
@@ -966,6 +979,7 @@ std::string VersionEdit::DebugJSON(int edit_num, bool hex_key) const {
        // permanent
        jw << "Temperature" << static_cast<int>(f.temperature);
      }
+      jw << "TailSize" << f.tail_size;
      jw.EndArrayedObject();
    }

@@ -90,6 +90,7 @@ enum NewFileCustomTag : uint32_t {
  kUniqueId = 12,
  kEpochNumber = 13,
  kCompensatedRangeDeletionSize = 14,
+  kTailSize = 15,

  // If this bit for the custom tag is set, opening DB should fail if
  // we don't know this field.
@@ -238,6 +239,10 @@ struct FileMetaData {
  // SST unique id
  UniqueId64x2 unique_id{};

+  // Size of the "tail" part of a SST file
+  // "Tail" refers to all blocks after data blocks till the end of the SST file
+  uint64_t tail_size = 0;
+
  FileMetaData() = default;

  FileMetaData(uint64_t file, uint32_t file_path_id, uint64_t file_size,
@@ -249,7 +254,8 @@ struct FileMetaData {
               uint64_t _epoch_number, const std::string& _file_checksum,
               const std::string& _file_checksum_func_name,
               UniqueId64x2 _unique_id,
-               const uint64_t _compensated_range_deletion_size)
+               const uint64_t _compensated_range_deletion_size,
+               uint64_t _tail_size)
      : fd(file, file_path_id, file_size, smallest_seq, largest_seq),
        smallest(smallest_key),
        largest(largest_key),
@@ -262,7 +268,8 @@ struct FileMetaData {
        epoch_number(_epoch_number),
        file_checksum(_file_checksum),
        file_checksum_func_name(_file_checksum_func_name),
-        unique_id(std::move(_unique_id)) {
+        unique_id(std::move(_unique_id)),
+        tail_size(_tail_size) {
    TEST_SYNC_POINT_CALLBACK("FileMetaData::FileMetaData", this);
  }

@@ -446,7 +453,8 @@ class VersionEdit {
               uint64_t epoch_number, const std::string& file_checksum,
               const std::string& file_checksum_func_name,
               const UniqueId64x2& unique_id,
-               const uint64_t compensated_range_deletion_size) {
+               const uint64_t compensated_range_deletion_size,
+               uint64_t tail_size) {
    assert(smallest_seqno <= largest_seqno);
    new_files_.emplace_back(
        level,
@@ -455,7 +463,7 @@ class VersionEdit {
                     temperature, oldest_blob_file_number, oldest_ancester_time,
                     file_creation_time, epoch_number, file_checksum,
                     file_checksum_func_name, unique_id,
-                     compensated_range_deletion_size));
+                     compensated_range_deletion_size, tail_size));
    if (!HasLastSequence() || largest_seqno > GetLastSequence()) {
      SetLastSequence(largest_seqno);
    }
@@ -566,13 +566,13 @@ Status VersionEditHandler::LoadTables(ColumnFamilyData* cfd,
  assert(builder_iter->second != nullptr);
  VersionBuilder* builder = builder_iter->second->version_builder();
  assert(builder);
+  const MutableCFOptions* moptions = cfd->GetLatestMutableCFOptions();
  Status s = builder->LoadTableHandlers(
      cfd->internal_stats(),
      version_set_->db_options_->max_file_opening_threads,
      prefetch_index_and_filter_in_cache, is_initial_load,
-      cfd->GetLatestMutableCFOptions()->prefix_extractor,
-      MaxFileSizeForL0MetaPin(*cfd->GetLatestMutableCFOptions()),
-      read_options_);
+      moptions->prefix_extractor, MaxFileSizeForL0MetaPin(*moptions),
+      read_options_, moptions->block_protection_bytes_per_key);
  if ((s.IsPathNotFound() || s.IsCorruption()) && no_error_if_files_missing_) {
    s = Status::OK();
  }
@@ -812,16 +812,16 @@ Status VersionEditHandlerPointInTime::MaybeCreateVersion(
      assert(builder);
    }

+    const MutableCFOptions* cf_opts_ptr = cfd->GetLatestMutableCFOptions();
    auto* version = new Version(cfd, version_set_, version_set_->file_options_,
-                                *cfd->GetLatestMutableCFOptions(), io_tracer_,
+                                *cf_opts_ptr, io_tracer_,
                                version_set_->current_version_number_++,
                                epoch_number_requirement_);
    s = builder->LoadTableHandlers(
        cfd->internal_stats(),
        version_set_->db_options_->max_file_opening_threads, false, true,
-        cfd->GetLatestMutableCFOptions()->prefix_extractor,
-        MaxFileSizeForL0MetaPin(*cfd->GetLatestMutableCFOptions()),
-        read_options_);
+        cf_opts_ptr->prefix_extractor, MaxFileSizeForL0MetaPin(*cf_opts_ptr),
+        read_options_, cf_opts_ptr->block_protection_bytes_per_key);
    if (!s.ok()) {
      delete version;
      if (s.IsCorruption()) {
@@ -45,7 +45,7 @@ TEST_F(VersionEditTest, EncodeDecode) {
                 kBig + 500 + i, kBig + 600 + i, false, Temperature::kUnknown,
                 kInvalidBlobFileNumber, 888, 678,
                 kBig + 300 + i /* epoch_number */, "234", "crc32c",
-                 kNullUniqueId64x2, 0);
+                 kNullUniqueId64x2, 0, 0);
    edit.DeleteFile(4, kBig + 700 + i);
  }

@@ -65,24 +65,24 @@ TEST_F(VersionEditTest, EncodeDecodeNewFile4) {
               kBig + 600, true, Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               300 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  edit.AddFile(4, 301, 3, 100, InternalKey("foo", kBig + 501, kTypeValue),
               InternalKey("zoo", kBig + 601, kTypeDeletion), kBig + 501,
               kBig + 601, false, Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               301 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  edit.AddFile(5, 302, 0, 100, InternalKey("foo", kBig + 502, kTypeValue),
               InternalKey("zoo", kBig + 602, kTypeDeletion), kBig + 502,
               kBig + 602, true, Temperature::kUnknown, kInvalidBlobFileNumber,
               666, 888, 302 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  edit.AddFile(5, 303, 0, 100, InternalKey("foo", kBig + 503, kTypeBlobIndex),
               InternalKey("zoo", kBig + 603, kTypeBlobIndex), kBig + 503,
               kBig + 603, true, Temperature::kUnknown, 1001,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               303 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  edit.DeleteFile(4, 700);

@@ -123,12 +123,12 @@ TEST_F(VersionEditTest, ForwardCompatibleNewFile4) {
               kBig + 600, true, Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               300 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  edit.AddFile(4, 301, 3, 100, InternalKey("foo", kBig + 501, kTypeValue),
               InternalKey("zoo", kBig + 601, kTypeDeletion), kBig + 501,
               kBig + 601, false, Temperature::kUnknown, kInvalidBlobFileNumber,
               686, 868, 301 /* epoch_number */, "234", "crc32c",
-               kNullUniqueId64x2, 0);
+               kNullUniqueId64x2, 0, 0);
  edit.DeleteFile(4, 700);

  edit.SetComparatorName("foo");
@@ -177,7 +177,7 @@ TEST_F(VersionEditTest, NewFile4NotSupportedField) {
               kBig + 600, true, Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               300 /* epoch_number */, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);

  edit.SetComparatorName("foo");
  edit.SetLogNumber(kBig + 100);
@@ -208,7 +208,7 @@ TEST_F(VersionEditTest, EncodeEmptyFile) {
               Temperature::kUnknown, kInvalidBlobFileNumber,
               kUnknownOldestAncesterTime, kUnknownFileCreationTime,
               1 /*epoch_number*/, kUnknownFileChecksum,
-               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0);
+               kUnknownFileChecksumFuncName, kNullUniqueId64x2, 0, 0);
  std::string buffer;
  ASSERT_TRUE(!edit.EncodeTo(&buffer));
 }
@@ -37,6 +37,7 @@
 #include "db/pinned_iterators_manager.h"
 #include "db/table_cache.h"
 #include "db/version_builder.h"
+#include "db/version_edit.h"
 #include "db/version_edit_handler.h"
 #include "table/compaction_merging_iterator.h"

@@ -941,7 +942,7 @@ class LevelIterator final : public InternalIterator {
      const std::shared_ptr<const SliceTransform>& prefix_extractor,
      bool should_sample, HistogramImpl* file_read_hist,
      TableReaderCaller caller, bool skip_filters, int level,
-      RangeDelAggregator* range_del_agg,
+      uint8_t block_protection_bytes_per_key, RangeDelAggregator* range_del_agg,
      const std::vector<AtomicCompactionUnitBoundary>* compaction_boundaries =
          nullptr,
      bool allow_unprepared_value = false,
@@ -964,6 +965,7 @@ class LevelIterator final : public InternalIterator {
        pinned_iters_mgr_(nullptr),
        compaction_boundaries_(compaction_boundaries),
        is_next_read_sequential_(false),
+        block_protection_bytes_per_key_(block_protection_bytes_per_key),
        range_tombstone_iter_(nullptr),
        to_return_sentinel_(false) {
    // Empty level is not supported.
@@ -1107,7 +1109,8 @@ class LevelIterator final : public InternalIterator {
        nullptr /* don't need reference to table */, file_read_hist_, caller_,
        /*arena=*/nullptr, skip_filters_, level_,
        /*max_file_size_for_l0_meta_pin=*/0, smallest_compaction_key,
-        largest_compaction_key, allow_unprepared_value_, range_tombstone_iter_);
+        largest_compaction_key, allow_unprepared_value_,
+        block_protection_bytes_per_key_, range_tombstone_iter_);
  }

  // Check if current file being fully within iterate_lower_bound.
@@ -1154,6 +1157,8 @@ class LevelIterator final : public InternalIterator {

  bool is_next_read_sequential_;

+  uint8_t block_protection_bytes_per_key_;
+
  // This is set when this level iterator is used under a merging iterator
  // that processes range tombstones. range_tombstone_iter_ points to where the
  // merging iterator stores the range tombstones iterator for this level. When
@@ -1535,6 +1540,7 @@ Status Version::GetTableProperties(const ReadOptions& read_options,
  auto ioptions = cfd_->ioptions();
  Status s = table_cache->GetTableProperties(
      file_options_, read_options, cfd_->internal_comparator(), *file_meta, tp,
+      mutable_cf_options_.block_protection_bytes_per_key,
      mutable_cf_options_.prefix_extractor, true /* no io */);
  if (s.ok()) {
    return s;
@@ -1621,6 +1627,7 @@ Status Version::TablesRangeTombstoneSummary(int max_entries_to_print,

      Status s = table_cache->GetRangeTombstoneIterator(
          read_options, cfd_->internal_comparator(), *file_meta,
+          cfd_->GetLatestMutableCFOptions()->block_protection_bytes_per_key,
          &tombstone_iter);
      if (!s.ok()) {
        return s;
@@ -1739,6 +1746,7 @@ size_t Version::GetMemoryUsageByTableReaders(const ReadOptions& read_options) {
      total_usage += cfd_->table_cache()->GetMemoryUsageByTableReader(
          file_options_, read_options, cfd_->internal_comparator(),
          *file_level.files[i].file_metadata,
+          mutable_cf_options_.block_protection_bytes_per_key,
          mutable_cf_options_.prefix_extractor);
    }
  }
@@ -1818,6 +1826,49 @@ uint64_t Version::GetSstFilesSize() {
  return sst_files_size;
 }

+void Version::GetSstFilesBoundaryKeys(Slice* smallest_user_key,
+                                      Slice* largest_user_key) {
+  smallest_user_key->clear();
+  largest_user_key->clear();
+  bool initialized = false;
+  const Comparator* ucmp = storage_info_.user_comparator_;
+  for (int level = 0; level < cfd_->NumberLevels(); level++) {
+    if (storage_info_.LevelFiles(level).size() == 0) {
+      continue;
+    }
+    if (level == 0) {
+      // we need to consider all files on level 0
+      for (const auto& file : storage_info_.LevelFiles(level)) {
+        const Slice& start_user_key = file->smallest.user_key();
+        if (!initialized ||
+            ucmp->Compare(start_user_key, *smallest_user_key) < 0) {
+          *smallest_user_key = start_user_key;
+        }
+        const Slice& end_user_key = file->largest.user_key();
+        if (!initialized ||
+            ucmp->Compare(end_user_key, *largest_user_key) > 0) {
+          *largest_user_key = end_user_key;
+        }
+        initialized = true;
+      }
+    } else {
+      // we only need to consider the first and last file
+      const Slice& start_user_key =
+          storage_info_.LevelFiles(level)[0]->smallest.user_key();
+      if (!initialized ||
+          ucmp->Compare(start_user_key, *smallest_user_key) < 0) {
+        *smallest_user_key = start_user_key;
+      }
+      const Slice& end_user_key =
+          storage_info_.LevelFiles(level).back()->largest.user_key();
+      if (!initialized || ucmp->Compare(end_user_key, *largest_user_key) > 0) {
+        *largest_user_key = end_user_key;
+      }
+      initialized = true;
+    }
+  }
+}
+
 void Version::GetCreationTimeOfOldestFile(uint64_t* creation_time) {
  uint64_t oldest_time = std::numeric_limits<uint64_t>::max();
  for (int level = 0; level < storage_info_.num_non_empty_levels_; level++) {
@@ -1848,6 +1899,7 @@ InternalIterator* Version::TEST_GetLevelIterator(
      mutable_cf_options_.prefix_extractor, should_sample_file_read(),
      cfd_->internal_stats()->GetFileReadHist(level),
      TableReaderCaller::kUserIterator, IsFilterSkipped(level), level,
+      mutable_cf_options_.block_protection_bytes_per_key,
      nullptr /* range_del_agg */, nullptr /* compaction_boundaries */,
      allow_unprepared_value, &tombstone_iter_ptr);
  if (read_options.ignore_range_deletions) {
@@ -1946,7 +1998,7 @@ void Version::AddIteratorsForLevel(const ReadOptions& read_options,
          /*skip_filters=*/false, /*level=*/0, max_file_size_for_l0_meta_pin_,
          /*smallest_compaction_key=*/nullptr,
          /*largest_compaction_key=*/nullptr, allow_unprepared_value,
-          &tombstone_iter);
+          mutable_cf_options_.block_protection_bytes_per_key, &tombstone_iter);
      if (read_options.ignore_range_deletions) {
        merge_iter_builder->AddIterator(table_iter);
      } else {
@@ -1975,8 +2027,10 @@ void Version::AddIteratorsForLevel(const ReadOptions& read_options,
        mutable_cf_options_.prefix_extractor, should_sample_file_read(),
        cfd_->internal_stats()->GetFileReadHist(level),
        TableReaderCaller::kUserIterator, IsFilterSkipped(level), level,
-        /*range_del_agg=*/nullptr, /*compaction_boundaries=*/nullptr,
-        allow_unprepared_value, &tombstone_iter_ptr);
+        mutable_cf_options_.block_protection_bytes_per_key,
+        /*range_del_agg=*/nullptr,
+        /*compaction_boundaries=*/nullptr, allow_unprepared_value,
+        &tombstone_iter_ptr);
    if (read_options.ignore_range_deletions) {
      merge_iter_builder->AddIterator(level_iter);
    } else {
@@ -2019,7 +2073,8 @@ Status Version::OverlapWithLevelIterator(const ReadOptions& read_options,
          /*skip_filters=*/false, /*level=*/0, max_file_size_for_l0_meta_pin_,
          /*smallest_compaction_key=*/nullptr,
          /*largest_compaction_key=*/nullptr,
-          /*allow_unprepared_value=*/false));
+          /*allow_unprepared_value=*/false,
+          mutable_cf_options_.block_protection_bytes_per_key));
      status = OverlapWithIterator(ucmp, smallest_user_key, largest_user_key,
                                   iter.get(), overlap);
      if (!status.ok() || *overlap) {
@@ -2034,7 +2089,8 @@ Status Version::OverlapWithLevelIterator(const ReadOptions& read_options,
        mutable_cf_options_.prefix_extractor, should_sample_file_read(),
        cfd_->internal_stats()->GetFileReadHist(level),
        TableReaderCaller::kUserIterator, IsFilterSkipped(level), level,
-        &range_del_agg));
+        mutable_cf_options_.block_protection_bytes_per_key, &range_del_agg,
+        nullptr, false));
    status = OverlapWithIterator(ucmp, smallest_user_key, largest_user_key,
                                 iter.get(), overlap);
  }
@@ -2333,7 +2389,8 @@ void Version::Get(const ReadOptions& read_options, const LookupKey& k,
    StopWatchNano timer(clock_, timer_enabled /* auto_start */);
    *status = table_cache_->Get(
        read_options, *internal_comparator(), *f->file_metadata, ikey,
-        &get_context, mutable_cf_options_.prefix_extractor,
+        &get_context, mutable_cf_options_.block_protection_bytes_per_key,
+        mutable_cf_options_.prefix_extractor,
        cfd_->internal_stats()->GetFileReadHist(fp.GetHitFileLevel()),
        IsFilterSkipped(static_cast<int>(fp.GetHitFileLevel()),
                        fp.IsHitFileLastInLevel()),
@@ -2578,7 +2635,8 @@ void Version::MultiGet(const ReadOptions& read_options, MultiGetRange* range,
                read_options, *internal_comparator(), *f->file_metadata,
                mutable_cf_options_.prefix_extractor,
                cfd_->internal_stats()->GetFileReadHist(fp.GetHitFileLevel()),
-                fp.GetHitFileLevel(), &file_range, &table_handle);
+                fp.GetHitFileLevel(), &file_range, &table_handle,
+                mutable_cf_options_.block_protection_bytes_per_key);
            skip_range_deletions = true;
            if (status.ok()) {
              skip_filters = true;
@@ -2768,7 +2826,8 @@ Status Version::ProcessBatch(
          read_options, *internal_comparator(), *f->file_metadata,
          mutable_cf_options_.prefix_extractor,
          cfd_->internal_stats()->GetFileReadHist(fp.GetHitFileLevel()),
-          fp.GetHitFileLevel(), &file_range, &table_handle);
+          fp.GetHitFileLevel(), &file_range, &table_handle,
+          mutable_cf_options_.block_protection_bytes_per_key);
      if (status.ok()) {
        skip_filters = true;
        skip_range_deletions = true;
@@ -3267,6 +3326,55 @@ uint32_t GetExpiredTtlFilesCount(const ImmutableOptions& ioptions,
  }
  return ttl_expired_files_count;
 }
+
+bool ShouldChangeFileTemperature(const ImmutableOptions& ioptions,
+                                 const MutableCFOptions& mutable_cf_options,
+                                 const std::vector<FileMetaData*>& files) {
+  const std::vector<FileTemperatureAge>& ages =
+      mutable_cf_options.compaction_options_fifo
+          .file_temperature_age_thresholds;
+  if (ages.empty()) {
+    return false;
+  }
+  if (files.empty()) {
+    return false;
+  }
+  int64_t _current_time;
+  auto status = ioptions.clock->GetCurrentTime(&_current_time);
+  const uint64_t current_time = static_cast<uint64_t>(_current_time);
+  // We use oldest_ancestor_time of a file to be the estimate age of
+  // the file just older than it. This is the same logic used in
+  // FIFOCompactionPicker::PickTemperatureChangeCompaction().
+  if (status.ok() && current_time >= ages[0].age) {
+    uint64_t create_time_threshold = current_time - ages[0].age;
+    Temperature target_temp;
+    assert(files.size() >= 1);
+    for (size_t index = files.size() - 1; index >= 1; --index) {
+      FileMetaData* cur_file = files[index];
+      FileMetaData* prev_file = files[index - 1];
+      if (!cur_file->being_compacted) {
+        uint64_t oldest_ancestor_time = prev_file->TryGetOldestAncesterTime();
+        if (oldest_ancestor_time == kUnknownOldestAncesterTime) {
+          return false;
+        }
+        if (oldest_ancestor_time > create_time_threshold) {
+          return false;
+        }
+        target_temp = ages[0].temperature;
+        for (size_t i = 1; i < ages.size(); ++i) {
+          if (current_time >= ages[i].age &&
+              oldest_ancestor_time <= current_time - ages[i].age) {
+            target_temp = ages[i].temperature;
+          }
+        }
+        if (cur_file->temperature != target_temp) {
+          return true;
+        }
+      }
+    }
+  }
+  return false;
+}
 }  // anonymous namespace

 void VersionStorageInfo::ComputeCompactionScore(
@@ -3324,22 +3432,25 @@ void VersionStorageInfo::ComputeCompactionScore(
      if (compaction_style_ == kCompactionStyleFIFO) {
        score = static_cast<double>(total_size) /
                mutable_cf_options.compaction_options_fifo.max_table_files_size;
-        if (mutable_cf_options.compaction_options_fifo.allow_compaction ||
-            mutable_cf_options.compaction_options_fifo.age_for_warm > 0) {
-          // Warm tier move can happen at any time. It's too expensive to
-          // check very file's timestamp now. For now, just trigger it
-          // slightly more frequently than FIFO compaction so that this
-          // happens first.
+        if (score < 1 &&
+            mutable_cf_options.compaction_options_fifo.allow_compaction) {
          score = std::max(
              static_cast<double>(num_sorted_runs) /
                  mutable_cf_options.level0_file_num_compaction_trigger,
              score);
        }
-        if (mutable_cf_options.ttl > 0) {
-          score = std::max(
-              static_cast<double>(GetExpiredTtlFilesCount(
-                  immutable_options, mutable_cf_options, files_[level])),
-              score);
+        if (score < 1 && mutable_cf_options.ttl > 0) {
+          score =
+              std::max(static_cast<double>(GetExpiredTtlFilesCount(
+                           immutable_options, mutable_cf_options, files_[0])),
+                       score);
+        }
+        if (score < 1 &&
+            ShouldChangeFileTemperature(immutable_options, mutable_cf_options,
+                                        files_[0])) {
+          // For FIFO, just need a large enough score to trigger compaction.
+          const double kScoreForNeedCompaction = 1.1;
+          score = kScoreForNeedCompaction;
        }
      } else {
        score = static_cast<double>(num_sorted_runs) /
@@ -5217,7 +5328,8 @@ Status VersionSet::ProcessManifestWrites(
            true /* prefetch_index_and_filter_in_cache */,
            false /* is_initial_load */,
            mutable_cf_options_ptrs[i]->prefix_extractor,
-            MaxFileSizeForL0MetaPin(*mutable_cf_options_ptrs[i]), read_options);
+            MaxFileSizeForL0MetaPin(*mutable_cf_options_ptrs[i]), read_options,
+            mutable_cf_options_ptrs[i]->block_protection_bytes_per_key);
        if (!s.ok()) {
          if (db_options_->paranoid_checks) {
            break;
@@ -6337,7 +6449,7 @@ Status VersionSet::WriteCurrentStateToManifest(
                       f->oldest_blob_file_number, f->oldest_ancester_time,
                       f->file_creation_time, f->epoch_number, f->file_checksum,
                       f->file_checksum_func_name, f->unique_id,
-                       f->compensated_range_deletion_size);
+                       f->compensated_range_deletion_size, f->tail_size);
        }
      }

@@ -6553,10 +6665,11 @@ uint64_t VersionSet::ApproximateOffsetOf(const ReadOptions& read_options,
    // "key" falls in the range for this table.  Add the
    // approximate offset of "key" within the table.
    TableCache* table_cache = v->cfd_->table_cache();
+    const MutableCFOptions& cf_opts = v->GetMutableCFOptions();
    if (table_cache != nullptr) {
      result = table_cache->ApproximateOffsetOf(
          read_options, key, *f.file_metadata, caller, icmp,
-          v->GetMutableCFOptions().prefix_extractor);
+          cf_opts.block_protection_bytes_per_key, cf_opts.prefix_extractor);
    }
  }
  return result;
@@ -6596,9 +6709,10 @@ uint64_t VersionSet::ApproximateSize(const ReadOptions& read_options,
  if (table_cache == nullptr) {
    return 0;
  }
+  const MutableCFOptions& cf_opts = v->GetMutableCFOptions();
  return table_cache->ApproximateSize(
      read_options, start, end, *f.file_metadata, caller, icmp,
-      v->GetMutableCFOptions().prefix_extractor);
+      cf_opts.block_protection_bytes_per_key, cf_opts.prefix_extractor);
 }

 void VersionSet::RemoveLiveFiles(
@@ -6757,6 +6871,7 @@ InternalIterator* VersionSet::MakeInputIterator(
              /*smallest_compaction_key=*/nullptr,
              /*largest_compaction_key=*/nullptr,
              /*allow_unprepared_value=*/false,
+              c->mutable_cf_options()->block_protection_bytes_per_key,
              /*range_del_iter=*/&range_tombstone_iter);
          range_tombstones.emplace_back(range_tombstone_iter, nullptr);
        }
@@ -6770,8 +6885,9 @@ InternalIterator* VersionSet::MakeInputIterator(
            /*should_sample=*/false,
            /*no per level latency histogram=*/nullptr,
            TableReaderCaller::kCompaction, /*skip_filters=*/false,
-            /*level=*/static_cast<int>(c->level(which)), range_del_agg,
-            c->boundaries(which), false, &tombstone_iter_ptr);
+            /*level=*/static_cast<int>(c->level(which)),
+            c->mutable_cf_options()->block_protection_bytes_per_key,
+            range_del_agg, c->boundaries(which), false, &tombstone_iter_ptr);
        range_tombstones.emplace_back(nullptr, tombstone_iter_ptr);
      }
    }
@@ -7008,7 +7124,8 @@ Status VersionSet::VerifyFileMetadata(const ReadOptions& read_options,
    TableCache::TypedHandle* handle = nullptr;
    FileMetaData meta_copy = meta;
    status = table_cache->FindTable(
-        read_options, file_opts, *icmp, meta_copy, &handle, pe,
+        read_options, file_opts, *icmp, meta_copy, &handle,
+        cf_opts->block_protection_bytes_per_key, pe,
        /*no_io=*/false, /*record_read_stats=*/true,
        internal_stats->GetFileReadHist(level), false, level,
        /*prefetch_index_and_filter_in_cache*/ false, max_sz_for_l0_meta_pin,
@@ -992,6 +992,9 @@ class Version {

  void GetColumnFamilyMetaData(ColumnFamilyMetaData* cf_meta);

+  void GetSstFilesBoundaryKeys(Slice* smallest_user_key,
+                               Slice* largest_user_key);
+
  uint64_t GetSstFilesSize();

  // Retrieves the file_creation_time of the oldest file in the DB.