ocfs2: adjust enabling place for la window

Patch series "improve write IO performance when fragmentation is high",
v6.


This patch (of 4):

After introducing gd->bg_contig_free_bits, the code path
'ocfs2_cluster_group_search() => ocfs2_local_alloc_seen_free_bits()'
becomes death when all the gd->bg_contig_free_bits are set to the correct
value. This patch relocates ocfs2_local_alloc_seen_free_bits() to a more
appropriate location. (The new place being ocfs2_block_group_set_bits().)

In ocfs2_local_alloc_seen_free_bits(), the scope of the spin-lock has been
adjusted to reduce meaningless lock races. e.g: when userspace creates &
deletes 1 cluster_size files in parallel, acquiring the spin-lock in
ocfs2_local_alloc_seen_free_bits() is totally pointless and impedes IO
performance.

Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Heming Zhao <[e-mail]>
Reviewed-by: Joseph Qi <[e-mail]>
Cc: Mark Fasheh <[e-mail]>
Cc: Joel Becker <[e-mail]>
Cc: Junxiao Bi <[e-mail]>
Cc: Changwei Ge <[e-mail]>
Cc: Gang He <[e-mail]>
Cc: Jun Piao <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>

fs/ocfs2/localalloc.c

···
 void ocfs2_local_alloc_seen_free_bits(struct ocfs2_super *osb,
 				      unsigned int num_clusters)
 {
-	spin_lock(&osb->osb_lock);
-	if (osb->local_alloc_state == OCFS2_LA_DISABLED ||
-	    osb->local_alloc_state == OCFS2_LA_THROTTLED)
-		if (num_clusters >= osb->local_alloc_default_bits) {
+	if (num_clusters >= osb->local_alloc_default_bits) {
+		spin_lock(&osb->osb_lock);
+		if (osb->local_alloc_state == OCFS2_LA_DISABLED ||
+		    osb->local_alloc_state == OCFS2_LA_THROTTLED) {
 			cancel_delayed_work(&osb->la_enable_wq);
 			osb->local_alloc_state = OCFS2_LA_ENABLED;
 		}
-	spin_unlock(&osb->osb_lock);
+		spin_unlock(&osb->osb_lock);
+	}
 }
 
 void ocfs2_la_enable_worker(struct work_struct *work)

fs/ocfs2/suballoc.c

···
 	int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
 	unsigned int start = bit_off + num_bits;
 	u16 contig_bits;
+	struct ocfs2_super *osb = OCFS2_SB(alloc_inode->i_sb);
 
 	/* All callers get the descriptor via
 	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
···
 		if (contig_bits > max_contig_bits)
 			max_contig_bits = contig_bits;
 		bg->bg_contig_free_bits = cpu_to_le16(max_contig_bits);
+		ocfs2_local_alloc_seen_free_bits(osb, max_contig_bits);
 	} else {
 		bg->bg_contig_free_bits = 0;
 	}
···
 		 * of bits. */
 		if (min_bits <= res->sr_bits)
 			search = 0; /* success */
-		else if (res->sr_bits) {
-			/*
-			 * Don't show bits which we'll be returning
-			 * for allocation to the local alloc bitmap.
-			 */
-			ocfs2_local_alloc_seen_free_bits(osb, res->sr_bits);
-		}
 	}
 
 	return search;