[sql] Use memory-mapped I/O for sql::Connection.

sql::Connection::Open*() uses PRAGMA mmap_size to enable SQLite's
memory-mapped I/O.  Additionally instrument to flush dirty pages from
the page cache after writes.

BUG=489784,533682

Review URL: https://codereview.chromium.org/1349863003

Cr-Commit-Position: refs/heads/master@{#350362}

4 files changed, 194 insertions, 0 deletions

diff --git a/sql/connection.cc b/sql/connection.cc
index 300e5e6..0b15387 100644
--- a/sql/connection.cc
+++ b/sql/connection.cc
@@ -466,6 +466,73 @@ void Connection::Preload() {
   }
 }
 
+// SQLite keeps unused pages associated with a connection in a cache.  It asks
+// the cache for pages by an id, and if the page is present and the database is
+// unchanged, it considers the content of the page valid and doesn't read it
+// from disk.  When memory-mapped I/O is enabled, on read SQLite uses page
+// structures created from the memory map data before consulting the cache.  On
+// write SQLite creates a new in-memory page structure, copies the data from the
+// memory map, and later writes it, releasing the updated page back to the
+// cache.
+//
+// This means that in memory-mapped mode, the contents of the cached pages are
+// not re-used for reads, but they are re-used for writes if the re-written page
+// is still in the cache. The implementation of sqlite3_db_release_memory() as
+// of SQLite 3.8.7.4 frees all pages from pcaches associated with the
+// connection, so it should free these pages.
+//
+// Unfortunately, the zero page is also freed.  That page is never accessed
+// using memory-mapped I/O, and the cached copy can be re-used after verifying
+// the file change counter on disk.  Also, fresh pages from cache receive some
+// pager-level initialization before they can be used.  Since the information
+// involved will immediately be accessed in various ways, it is unclear if the
+// additional overhead is material, or just moving processor cache effects
+// around.
+//
+// TODO(shess): It would be better to release the pages immediately when they
+// are no longer needed.  This would basically happen after SQLite commits a
+// transaction.  I had implemented a pcache wrapper to do this, but it involved
+// layering violations, and it had to be setup before any other sqlite call,
+// which was brittle.  Also, for large files it would actually make sense to
+// maintain the existing pcache behavior for blocks past the memory-mapped
+// segment.  I think drh would accept a reasonable implementation of the overall
+// concept for upstreaming to SQLite core.
+//
+// TODO(shess): Another possibility would be to set the cache size small, which
+// would keep the zero page around, plus some pre-initialized pages, and SQLite
+// can manage things.  The downside is that updates larger than the cache would
+// spill to the journal.  That could be compensated by setting cache_spill to
+// false.  The downside then is that it allows open-ended use of memory for
+// large transactions.
+//
+// TODO(shess): The TrimMemory() trick of bouncing the cache size would also
+// work.  There could be two prepared statements, one for cache_size=1 one for
+// cache_size=goal.
+void Connection::ReleaseCacheMemoryIfNeeded(bool implicit_change_performed) {
+  // If memory-mapping is not enabled, the page cache helps performance.
+  if (!mmap_enabled_)
+    return;
+
+  // On caller request, force the change comparison to fail.  Done before the
+  // transaction-nesting test so that the signal can carry to transaction
+  // commit.
+  if (implicit_change_performed)
+    --total_changes_at_last_release_;
+
+  // Cached pages may be re-used within the same transaction.
+  if (transaction_nesting())
+    return;
+
+  // If no changes have been made, skip flushing.  This allows the first page of
+  // the database to remain in cache across multiple reads.
+  const int total_changes = sqlite3_total_changes(db_);
+  if (total_changes == total_changes_at_last_release_)
+    return;
+
+  total_changes_at_last_release_ = total_changes;
+  sqlite3_db_release_memory(db_);
+}
+
 void Connection::TrimMemory(bool aggressively) {
   if (!db_)
     return;
@@ -776,6 +843,9 @@ bool Connection::CommitTransaction() {
   RecordCommitTime(delta);
   RecordOneEvent(EVENT_COMMIT);
 
+  // Release dirty cache pages after the transaction closes.
+  ReleaseCacheMemoryIfNeeded(false);
+
   return ret;
 }
 
@@ -866,6 +936,12 @@ int Connection::ExecuteAndReturnErrorCode(const char* sql) {
     const base::TimeDelta delta = Now() - before;
     RecordTimeAndChanges(delta, read_only);
   }
+
+  // Most calls to Execute() modify the database.  The main exceptions would be
+  // calls such as CREATE TABLE IF NOT EXISTS which could modify the database
+  // but sometimes don't.
+  ReleaseCacheMemoryIfNeeded(true);
+
   return rc;
 }
 
@@ -1230,6 +1306,18 @@ bool Connection::OpenInternal(const std::string& file_name,
   // secure_delete.
   ignore_result(Execute("PRAGMA journal_mode = TRUNCATE"));
 
+  // Enable memory-mapped access.  This value will be capped by
+  // SQLITE_MAX_MMAP_SIZE, which could be different between 32-bit and 64-bit
+  // platforms.
+  mmap_enabled_ = false;
+  if (!mmap_disabled_)
+    ignore_result(Execute("PRAGMA mmap_size = 268435456"));  // 256MB.
+  {
+    Statement s(GetUniqueStatement("PRAGMA mmap_size"));
+    if (s.Step() && s.ColumnInt64(0) > 0)
+      mmap_enabled_ = true;
+  }
+
   const base::TimeDelta kBusyTimeout =
     base::TimeDelta::FromSeconds(kBusyTimeoutSeconds);
 
@@ -1273,6 +1361,9 @@ void Connection::DoRollback() {
   RecordUpdateTime(delta);
   RecordOneEvent(EVENT_ROLLBACK);
 
+  // The cache may have been accumulating dirty pages for commit.
+  ReleaseCacheMemoryIfNeeded(false);
+
   needs_rollback_ = false;
 }
 
diff --git a/sql/connection.h b/sql/connection.h
index 19592d9..d456b6c 100644
--- a/sql/connection.h
+++ b/sql/connection.h
@@ -146,6 +146,9 @@ class SQL_EXPORT Connection {
   // other platforms.
   void set_restrict_to_user() { restrict_to_user_ = true; }
 
+  // Call to opt out of memory-mapped file I/O.
+  void set_mmap_disabled() { mmap_disabled_ = true; }
+
   // Set an error-handling callback.  On errors, the error number (and
   // statement, if available) will be passed to the callback.
   //
@@ -638,6 +641,12 @@ class SQL_EXPORT Connection {
     return clock_->Now();
   }
 
+  // Release page-cache memory if memory-mapped I/O is enabled and the database
+  // was changed.  Passing true for |implicit_change_performed| allows
+  // overriding the change detection for cases like DDL (CREATE, DROP, etc),
+  // which do not participate in the total-rows-changed tracking.
+  void ReleaseCacheMemoryIfNeeded(bool implicit_change_performed);
+
   // The actual sqlite database. Will be NULL before Init has been called or if
   // Init resulted in an error.
   sqlite3* db_;
@@ -679,6 +688,17 @@ class SQL_EXPORT Connection {
   // databases.
   bool poisoned_;
 
+  // |true| if SQLite memory-mapped I/O is not desired for this connection.
+  bool mmap_disabled_;
+
+  // |true| if SQLite memory-mapped I/O was enabled for this connection.
+  // Used by ReleaseCacheMemoryIfNeeded().
+  bool mmap_enabled_;
+
+  // Used by ReleaseCacheMemoryIfNeeded() to track if new changes have happened
+  // since memory was last released.
+  int total_changes_at_last_release_;
+
   ErrorCallback error_callback_;
 
   // Tag for auxiliary histograms.
diff --git a/sql/connection_unittest.cc b/sql/connection_unittest.cc
index 10a15a8..5412be1 100644
--- a/sql/connection_unittest.cc
+++ b/sql/connection_unittest.cc
@@ -4,10 +4,12 @@
 
 #include "base/bind.h"
 #include "base/files/file_util.h"
+#include "base/files/memory_mapped_file.h"
 #include "base/files/scoped_file.h"
 #include "base/files/scoped_temp_dir.h"
 #include "base/logging.h"
 #include "base/metrics/statistics_recorder.h"
+#include "base/strings/stringprintf.h"
 #include "base/test/histogram_tester.h"
 #include "sql/connection.h"
 #include "sql/correct_sql_test_base.h"
@@ -1298,4 +1300,80 @@ TEST_F(SQLConnectionTest, TimeUpdateTransaction) {
   EXPECT_EQ(0, samples->sum());
 }
 
+// Make sure that OS file writes to a mmap'ed file are reflected in the memory
+// mapping of a memory-mapped file.  Normally SQLite writes to memory-mapped
+// files using memcpy(), which should stay consistent.  Our SQLite is slightly
+// patched to mmap read only, then write using OS file writes.  If the
+// memory-mapped version doesn't reflect the OS file writes, SQLite's
+// memory-mapped I/O should be disabled on this platform.
+TEST_F(SQLConnectionTest, MmapTest) {
+  // Skip the test for platforms which don't enable memory-mapped I/O in SQLite,
+  // or which don't even support the pragma.  The former seems to apply to iOS,
+  // the latter to older iOS.
+  // TODO(shess): Disable test on iOS?  Disable on USE_SYSTEM_SQLITE?
+  {
+    sql::Statement s(db().GetUniqueStatement("PRAGMA mmap_size"));
+    if (!s.Step() || !s.ColumnInt64(0))
+      return;
+  }
+
+  // The test re-uses the database file to make sure it's representative of a
+  // SQLite file, but will be storing incompatible data.
+  db().Close();
+
+  const uint32 kFlags =
+      base::File::FLAG_OPEN|base::File::FLAG_READ|base::File::FLAG_WRITE;
+  char buf[4096];
+
+  // Create a file with a block of '0', a block of '1', and a block of '2'.
+  {
+    base::File f(db_path(), kFlags);
+    ASSERT_TRUE(f.IsValid());
+    memset(buf, '0', sizeof(buf));
+    ASSERT_EQ(f.Write(0*sizeof(buf), buf, sizeof(buf)), (int)sizeof(buf));
+
+    memset(buf, '1', sizeof(buf));
+    ASSERT_EQ(f.Write(1*sizeof(buf), buf, sizeof(buf)), (int)sizeof(buf));
+
+    memset(buf, '2', sizeof(buf));
+    ASSERT_EQ(f.Write(2*sizeof(buf), buf, sizeof(buf)), (int)sizeof(buf));
+  }
+
+  // mmap the file and verify that everything looks right.
+  {
+    base::MemoryMappedFile m;
+    ASSERT_TRUE(m.Initialize(db_path()));
+
+    memset(buf, '0', sizeof(buf));
+    ASSERT_EQ(0, memcmp(buf, m.data() + 0*sizeof(buf), sizeof(buf)));
+
+    memset(buf, '1', sizeof(buf));
+    ASSERT_EQ(0, memcmp(buf, m.data() + 1*sizeof(buf), sizeof(buf)));
+
+    memset(buf, '2', sizeof(buf));
+    ASSERT_EQ(0, memcmp(buf, m.data() + 2*sizeof(buf), sizeof(buf)));
+
+    // Scribble some '3' into the first page of the file, and verify that it
+    // looks the same in the memory mapping.
+    {
+      base::File f(db_path(), kFlags);
+      ASSERT_TRUE(f.IsValid());
+      memset(buf, '3', sizeof(buf));
+      ASSERT_EQ(f.Write(0*sizeof(buf), buf, sizeof(buf)), (int)sizeof(buf));
+    }
+    ASSERT_EQ(0, memcmp(buf, m.data() + 0*sizeof(buf), sizeof(buf)));
+
+    // Repeat with a single '4' in case page-sized blocks are different.
+    const size_t kOffset = 1*sizeof(buf) + 123;
+    ASSERT_NE('4', m.data()[kOffset]);
+    {
+      base::File f(db_path(), kFlags);
+      ASSERT_TRUE(f.IsValid());
+      buf[0] = '4';
+      ASSERT_EQ(f.Write(kOffset, buf, 1), 1);
+    }
+    ASSERT_EQ('4', m.data()[kOffset]);
+  }
+}
+
 }  // namespace
diff --git a/sql/statement.cc b/sql/statement.cc
index 09c95da..2a0faa7 100644
--- a/sql/statement.cc
+++ b/sql/statement.cc
@@ -110,6 +110,11 @@ void Statement::Reset(bool clear_bound_vars) {
       ref_->connection()->RecordOneEvent(Connection::EVENT_STATEMENT_SUCCESS);
   }
 
+  // Potentially release dirty cache pages if an autocommit statement made
+  // changes.
+  if (ref_->connection())
+    ref_->connection()->ReleaseCacheMemoryIfNeeded(false);
+
   succeeded_ = false;
   stepped_ = false;
 }