add prefetch to PosixRandomAccessFile in buffered io

Summary:
Every time after a compaction/flush finish, we issue user reads to put the table into block cache which includes a couple of IO that read footer, index blocks, meta block, etc. So we implement Prefetch here to reduce IO.
Closes https://github.com/facebook/rocksdb/pull/2196

Differential Revision: D4931782

Pulled By: lightmark

fbshipit-source-id: 5a13d58dcab209964352322217193bbf7ff78149

diff --git a/env/io_posix.cc b/env/io_posix.cc
index 41b1204027de7f46f895185101271fb8ad9d1166..68cb70d2388ef1784f0c44c3c43cd1d17af8c074 100644 (file)
--- a/env/io_posix.cc
+++ b/env/io_posix.cc
@@ -38,7 +38,7 @@
 namespace rocksdb {
 
 // A wrapper for fadvise, if the platform doesn't support fadvise,
-// it will simply return Status::NotSupport.
+// it will simply return 0.
 int Fadvise(int fd, off_t offset, size_t len, int advice) {
 #ifdef OS_LINUX
   return posix_fadvise(fd, offset, len, advice);
@@ -337,6 +337,26 @@ Status PosixRandomAccessFile::Read(uint64_t offset, size_t n, Slice* result,
   return s;
 }
 
+Status PosixRandomAccessFile::Prefetch(uint64_t offset, size_t n) {
+  Status s;
+  if (!use_direct_io()) {
+    ssize_t r = 0;
+#ifdef OS_LINUX
+    r = readahead(fd_, offset, n);
+#endif
+#ifdef OS_MACOSX
+    radvisory advice;
+    advice.ra_offset = static_cast<off_t>(offset);
+    advice.ra_count = static_cast<int>(n);
+    r = fcntl(fd_, F_RDADVISE, &advice);
+#endif
+    if (r == -1) {
+      s = IOError(filename_, errno);
+    }
+  }
+  return s;
+}
+
 #if defined(OS_LINUX) || defined(OS_MACOSX) || defined(OS_AIX)
 size_t PosixRandomAccessFile::GetUniqueId(char* id, size_t max_size) const {
   return PosixHelper::GetUniqueIdFromFile(fd_, id, max_size);

diff --git a/env/io_posix.h b/env/io_posix.h
index 351d6d7cb94ad5825fcb38ffa525ece32ef4fcd0..af59939a911aab639607d26f1b355f8b49081c8d 100644 (file)
--- a/env/io_posix.h
+++ b/env/io_posix.h
@@ -79,6 +79,9 @@ class PosixRandomAccessFile : public RandomAccessFile {
 
   virtual Status Read(uint64_t offset, size_t n, Slice* result,
                       char* scratch) const override;
+
+  virtual Status Prefetch(uint64_t offset, size_t n) override;
+
 #if defined(OS_LINUX) || defined(OS_MACOSX) || defined(OS_AIX)
   virtual size_t GetUniqueId(char* id, size_t max_size) const override;
 #endif