MultiCFSnapshot for NewIterators() API (#12573)

Summary:
As mentioned in https://github.com/facebook/rocksdb/issues/12561 and https://github.com/facebook/rocksdb/issues/12566 , `NewIterators()` API has not been providing consistent view of the db across multiple column families. This PR addresses it by utilizing `MultiCFSnapshot()` function which has been used for `MultiGet()` APIs. To be able to obtain the thread-local super version with ref, `sv_exclusive_access` parameter has been added to `MultiCFSnapshot()` so that we could call `GetReferencedSuperVersion()` or `GetAndRefSuperVersion()` depending on the param and support `Refresh()` API for MultiCfIterators

Pull Request resolved: https://github.com/facebook/rocksdb/pull/12573

Test Plan:
**Unit Tests Added**

```
./db_iterator_test --gtest_filter="*IteratorsConsistentView*"
```
```
./multi_cf_iterator_test -- --gtest_filter="*ConsistentView*"
```

**Performance Check**

Setup
```
make -j64 release
TEST_TMPDIR=/dev/shm/db_bench ./db_bench -benchmarks="filluniquerandom" -key_size=32 -value_size=512 -num=10000000 -compression_type=none
```

Run
```
TEST_TMPDIR=/dev/shm/db_bench ./db_bench -use_existing_db=1 -benchmarks="multireadrandom" -cache_size=10485760000
```
Before the change
```
DB path: [/dev/shm/db_bench/dbbench]
multireadrandom :       6.374 micros/op 156892 ops/sec 6.374 seconds 1000000 operations; (0 of 1000000 found)
```
After the change
```
DB path: [/dev/shm/db_bench/dbbench]
multireadrandom :       6.265 micros/op 159627 ops/sec 6.265 seconds 1000000 operations; (0 of 1000000 found)
```

Reviewed By: jowlyzhang

Differential Revision: D56444066

Pulled By: jaykorean

fbshipit-source-id: 327ce73c072da30c221e18d4f3389f49115b8f99

1 files changed, 17 insertions, 6 deletions

diff --git a/db/db_impl/db_impl.h b/db/db_impl/db_impl.h
index ae9a8d5a1..504d7ec60 100644
--- a/db/db_impl/db_impl.h
+++ b/db/db_impl/db_impl.h
@@ -2355,18 +2355,20 @@ class DBImpl : public DB {
 
   // A structure to contain ColumnFamilyData and the SuperVersion obtained for
   // the consistent view of DB
-  struct ColumnFamilyDataSuperVersionPair {
+  struct ColumnFamilySuperVersionPair {
+    ColumnFamilyHandleImpl* cfh;
     ColumnFamilyData* cfd;
 
     // SuperVersion for the column family obtained in a manner that ensures a
     // consistent view across all column families in the DB
     SuperVersion* super_version;
-    ColumnFamilyDataSuperVersionPair(ColumnFamilyHandle* column_family,
-                                     SuperVersion* sv)
-        : cfd(static_cast<ColumnFamilyHandleImpl*>(column_family)->cfd()),
+    ColumnFamilySuperVersionPair(ColumnFamilyHandle* column_family,
+                                 SuperVersion* sv)
+        : cfh(static_cast<ColumnFamilyHandleImpl*>(column_family)),
+          cfd(cfh->cfd()),
           super_version(sv) {}
 
-    ColumnFamilyDataSuperVersionPair() = default;
+    ColumnFamilySuperVersionPair() = default;
   };
 
   // A common function to obtain a consistent snapshot, which can be implicit
@@ -2380,9 +2382,17 @@ class DBImpl : public DB {
   // If callback is non-null, the callback is refreshed with the snapshot
   // sequence number
   //
+  // `extra_sv_ref` is used to indicate whether thread-local SuperVersion
+  // should be obtained with an extra ref (by GetReferencedSuperVersion()) or
+  // not (by GetAndRefSuperVersion()). For instance, point lookup like MultiGet
+  // does not require SuperVersion to be re-acquired throughout the entire
+  // invocation (no need extra ref), while MultiCfIterators may need the
+  // SuperVersion to be updated during Refresh() (requires extra ref).
+  //
   // `sv_from_thread_local` being set to false indicates that the SuperVersion
   // obtained from the ColumnFamilyData, whereas true indicates they are thread
   // local.
+  //
   // A non-OK status will be returned if for a column family that enables
   // user-defined timestamp feature, the specified `ReadOptions.timestamp`
   // attemps to read collapsed history.
@@ -2390,7 +2400,8 @@ class DBImpl : public DB {
   Status MultiCFSnapshot(const ReadOptions& read_options,
                          ReadCallback* callback,
                          IterDerefFuncType iter_deref_func, T* cf_list,
-                         SequenceNumber* snapshot, bool* sv_from_thread_local);
+                         bool extra_sv_ref, SequenceNumber* snapshot,
+                         bool* sv_from_thread_local);
 
   // The actual implementation of the batching MultiGet. The caller is expected
   // to have acquired the SuperVersion and pass in a snapshot sequence number