fs-writeback: unplug before cond_resched in writeback_sb_inodes

Commit 505a666 ("writeback: plug writeback in wb_writeback() and
writeback_inodes_wb()") has us holding a plug during writeback_sb_inodes,
which increases the merge rate when relatively contiguous small files
are written by the filesystem.  It helps both on flash and spindles.

For an fs_mark workload creating 4K files in parallel across 8 drives,
this commit improves performance ~9% more by unplugging before calling
cond_resched().  cond_resched() doesn't trigger an implicit unplug, so
explicitly getting the IO down to the device before scheduling reduces
latencies for anyone waiting on clean pages.

It also cuts down on how often we use kblockd to unplug, which means
less work bouncing from one workqueue to another.

Many more details about how we got here:

  https://lkml.org/lkml/2015/9/11/570

Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/fs-writeback.c

@@ -1481,14 +1481,29 @@ static long writeback_sb_inodes(struct super_block *sb,
 		wbc_detach_inode(&wbc);
 		work->nr_pages -= write_chunk - wbc.nr_to_write;
 		wrote += write_chunk - wbc.nr_to_write;
+
+		if (need_resched()) {
+			/*
+			 * We're trying to balance between building up a nice
+			 * long list of IOs to improve our merge rate, and
+			 * getting those IOs out quickly for anyone throttling
+			 * in balance_dirty_pages().  cond_resched() doesn't
+			 * unplug, so get our IOs out the door before we
+			 * give up the CPU.
+			 */
+			blk_flush_plug(current);
+			cond_resched();
+		}
+
+
 		spin_lock(&wb->list_lock);
 		spin_lock(&inode->i_lock);
 		if (!(inode->i_state & I_DIRTY_ALL))
 			wrote++;
 		requeue_inode(inode, wb, &wbc);
 		inode_sync_complete(inode);
 		spin_unlock(&inode->i_lock);
-		cond_resched_lock(&wb->list_lock);
+
 		/*
 		 * bail out to wb_writeback() often enough to check
 		 * background threshold and other termination conditions.