f2fs: issue small discard by LBA order

For small granularity discard which size is smaller than 64KB, if we
issue those kind of discards orderly by size, their IOs will be spread
into entire logical address, so that in FTL, L2P table will be updated
randomly, result bad wear rate in the table.

In this patch, we choose to issue small discard by LBA order, by this
way, we can expect that L2P table updates from adjacent discard IOs can
be merged in the cache, so it can reduce lifetime wearing of flash.

Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index dfc4054..c8f88f0 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -934,6 +934,7 @@
 	/* common policy */
 	dpolicy->type = discard_type;
 	dpolicy->sync = true;
+	dpolicy->ordered = false;
 	dpolicy->granularity = granularity;
 
 	dpolicy->max_requests = DEF_MAX_DISCARD_REQUEST;
@@ -945,6 +946,7 @@
 		dpolicy->max_interval = DEF_MAX_DISCARD_ISSUE_TIME;
 		dpolicy->io_aware = true;
 		dpolicy->sync = false;
+		dpolicy->ordered = true;
 		if (utilization(sbi) > DEF_DISCARD_URGENT_UTIL) {
 			dpolicy->granularity = 1;
 			dpolicy->max_interval = DEF_MIN_DISCARD_ISSUE_TIME;
@@ -1181,6 +1183,63 @@
 	return 0;
 }
 
+static unsigned int __issue_discard_cmd_orderly(struct f2fs_sb_info *sbi,
+					struct discard_policy *dpolicy)
+{
+	struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info;
+	struct discard_cmd *prev_dc = NULL, *next_dc = NULL;
+	struct rb_node **insert_p = NULL, *insert_parent = NULL;
+	struct discard_cmd *dc;
+	struct blk_plug plug;
+	unsigned int pos = dcc->next_pos;
+	unsigned int issued = 0;
+	bool io_interrupted = false;
+
+	mutex_lock(&dcc->cmd_lock);
+	dc = (struct discard_cmd *)f2fs_lookup_rb_tree_ret(&dcc->root,
+					NULL, pos,
+					(struct rb_entry **)&prev_dc,
+					(struct rb_entry **)&next_dc,
+					&insert_p, &insert_parent, true);
+	if (!dc)
+		dc = next_dc;
+
+	blk_start_plug(&plug);
+
+	while (dc) {
+		struct rb_node *node;
+
+		if (dc->state != D_PREP)
+			goto next;
+
+		if (dpolicy->io_aware && !is_idle(sbi)) {
+			io_interrupted = true;
+			break;
+		}
+
+		dcc->next_pos = dc->lstart + dc->len;
+		__submit_discard_cmd(sbi, dpolicy, dc);
+
+		if (++issued >= dpolicy->max_requests)
+			break;
+next:
+		node = rb_next(&dc->rb_node);
+		dc = rb_entry_safe(node, struct discard_cmd, rb_node);
+	}
+
+	blk_finish_plug(&plug);
+
+	if (!dc)
+		dcc->next_pos = 0;
+
+	mutex_unlock(&dcc->cmd_lock);
+
+	if (!issued && io_interrupted)
+		issued = -1;
+
+	return issued;
+}
+
 static int __issue_discard_cmd(struct f2fs_sb_info *sbi,
 					struct discard_policy *dpolicy)
 {
@@ -1194,6 +1253,10 @@
 	for (i = MAX_PLIST_NUM - 1; i >= 0; i--) {
 		if (i + 1 < dpolicy->granularity)
 			break;
+
+		if (i < DEFAULT_DISCARD_GRANULARITY && dpolicy->ordered)
+			return __issue_discard_cmd_orderly(sbi, dpolicy);
+
 		pend_list = &dcc->pend_list[i];
 
 		mutex_lock(&dcc->cmd_lock);
@@ -1754,6 +1817,7 @@
 	dcc->nr_discards = 0;
 	dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg;
 	dcc->undiscard_blks = 0;
+	dcc->next_pos = 0;
 	dcc->root = RB_ROOT;
 	dcc->rbtree_check = false;