Avoid atomic operation in MarkLocalBufferDirty().

anarazel · Alexander Korotkov · commit 3e485feba4c6 · 2016-07-11T11:33:31.000+03:00
The recent patch to make Pin/UnpinBuffer lockfree in the hot path (4835458), accidentally used pg_atomic_fetch_or_u32() in MarkLocalBufferDirty(). Other code operating on local buffers was careful to only use pg_atomic_read/write_u32 which just read/write from memory; to avoid unnecessary overhead. On its own that'd just make MarkLocalBufferDirty() slightly less efficient, but in addition InitLocalBuffers() doesn't call pg_atomic_init_u32() - thus the spinlock fallback for the atomic operations isn't initialized. That in turn caused, as reported by Tom, buildfarm animal gaur to fail. As those errors are actually useful against this type of error, continue to omit - intentionally this time - initialization of the atomic variable. In addition, add an explicit note about only using pg_atomic_read/write on local buffers's state to BufferDesc's description. Reported-By: Tom Lane Discussion: 1881.1460431476@sss.pgh.pa.us
diff --git a/src/backend/storage/buffer/localbuf.c b/src/backend/storage/buffer/localbuf.c
@@ -295,10 +295,14 @@ MarkLocalBufferDirty(Buffer buffer)
 
 	bufHdr = GetLocalBufferDescriptor(bufid);
 
-	buf_state = pg_atomic_fetch_or_u32(&bufHdr->state, BM_DIRTY);
+	buf_state = pg_atomic_read_u32(&bufHdr->state);
 
 	if (!(buf_state & BM_DIRTY))
 		pgBufferUsage.local_blks_dirtied++;
+
+	buf_state |= BM_DIRTY;
+
+	pg_atomic_write_u32(&bufHdr->state, buf_state);
 }
 
 /*
@@ -432,6 +436,13 @@ InitLocalBuffers(void)
 		 * is -1.)
 		 */
 		buf->buf_id = -i - 2;
+
+		/*
+		 * Intentionally do not initialize the buffer's atomic variable
+		 * (besides zeroing the underlying memory above). That way we get
+		 * errors on platforms without atomics, if somebody (re-)introduces
+		 * atomic operations for local buffers.
+		 */
 	}
 
 	/* Create the lookup hash table */
diff --git a/src/include/storage/buf_internals.h b/src/include/storage/buf_internals.h
@@ -165,8 +165,19 @@ typedef struct buftag
  * wait_backend_pid and setting flag bit BM_PIN_COUNT_WAITER.  At present,
  * there can be only one such waiter per buffer.
  *
+<<<<<<< HEAD
  * We use this same struct for local buffer headers, but the lock fields
  * are not used and not all of the flag bits are useful either.
+=======
+ * We use this same struct for local buffer headers, but the locks are not
+ * used and not all of the flag bits are useful either. To avoid unnecessary
+ * overhead, manipulations of the state field should be done without actual
+ * atomic operations (i.e. only pg_atomic_read/write).
+ *
+ * Be careful to avoid increasing the size of the struct when adding or
+ * reordering members.  Keeping it below 64 bytes (the most common CPU
+ * cache line size) is fairly important for performance.
+>>>>>>> 6b93fcd... Avoid atomic operation in MarkLocalBufferDirty().
  */
 typedef struct BufferDesc
 {
