Skip to contentSkip to conversation

cfq-iosched: Get rid of st->active

PATCH In review Gui Jianfeng 3 participants 3 messages Nov 30, 2010 Block layer

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 cfq-iosched: Get rid of st->active

When a cfq group is running, it won't be dequeued from service tree, so there's no need to store the active one in st->active. Just gid rid of it.

---
 block/cfq-iosched.c |   15 +--------------
 1 files changed, 1 insertions(+), 14 deletions(-)
block/cfq-iosched.c +1 -14
index 73a5862..e18d316 100644
@@ -87,7 +87,6 @@ struct cfq_rb_root {
 	unsigned count;
 	unsigned total_weight;
 	u64 min_vdisktime;
-	struct rb_node *active;
 };
 #define CFQ_RB_ROOT	(struct cfq_rb_root) { .rb = RB_ROOT, .left = NULL, \
 			.count = 0, .min_vdisktime = 0, }
@@ -563,11 +562,6 @@ static void update_min_vdisktime(struct cfq_rb_root *st)
 	u64 vdisktime = st->min_vdisktime;
 	struct cfq_group *cfqg;
 
-	if (st->active) {
-		cfqg = rb_entry_cfqg(st->active);
-		vdisktime = cfqg->vdisktime;
-	}
-
 	if (st->left) {
 		cfqg = rb_entry_cfqg(st->left);
 		vdisktime = min_vdisktime(vdisktime, cfqg->vdisktime);
@@ -894,9 +888,6 @@ cfq_group_service_tree_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
 {
 	struct cfq_rb_root *st = &cfqd->grp_service_tree;
 
-	if (st->active == &cfqg->rb_node)
-		st->active = NULL;
-
 	BUG_ON(cfqg->nr_cfqq < 1);
 	cfqg->nr_cfqq--;
 
@@ -1095,7 +1086,7 @@ static void cfq_put_cfqg(struct cfq_group *cfqg)
 	if (!atomic_dec_and_test(&cfqg->ref))
 		return;
 	for_each_cfqg_st(cfqg, i, j, st)
-		BUG_ON(!RB_EMPTY_ROOT(&st->rb) || st->active != NULL);
+		BUG_ON(!RB_EMPTY_ROOT(&st->rb));
 	kfree(cfqg);
 }
 
@@ -1687,9 +1678,6 @@ __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	if (cfqq == cfqd->active_queue)
 		cfqd->active_queue = NULL;
 
-	if (&cfqq->cfqg->rb_node == cfqd->grp_service_tree.active)
-		cfqd->grp_service_tree.active = NULL;
-
 	if (cfqd->active_cic) {
 		put_io_context(cfqd->active_cic->ioc);
 		cfqd->active_cic = NULL;
@@ -2199,7 +2187,6 @@ static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd)
 	if (RB_EMPTY_ROOT(&st->rb))
 		return NULL;
 	cfqg = cfq_rb_first_group(st);
-	st->active = &cfqg->rb_node;
 	update_min_vdisktime(st);
 	return cfqg;
 }
Signature · 1 line
1.6.5.2
Acked-by on v1 1/1
#1 lore raw
Quoted 5 lines · Gui Jianfeng

When a cfq group is running, it won't be dequeued from service tree, so there's no need to store the active one in st->active. Just gid rid of it.

Looks good to me. st->left and st->active seems to be same as we never dequeue the group while it is being serviced.

I think in previous implementations I had first dequeued the group while it is being serviced and in that case st->left and st->active will be different, hence this additional pointer. Later changed the implementation and now st->active should not be needed.

Thanks for the cleanup Gui.

Vivek

Quoted 69 lines
---
 block/cfq-iosched.c |   15 +--------------
 1 files changed, 1 insertions(+), 14 deletions(-)
block/cfq-iosched.c +1 -14
index 73a5862..e18d316 100644
@@ -87,7 +87,6 @@ struct cfq_rb_root {
 	unsigned count;
 	unsigned total_weight;
 	u64 min_vdisktime;
-	struct rb_node *active;
 };
 #define CFQ_RB_ROOT	(struct cfq_rb_root) { .rb = RB_ROOT, .left = NULL, \
 			.count = 0, .min_vdisktime = 0, }
@@ -563,11 +562,6 @@ static void update_min_vdisktime(struct cfq_rb_root *st)
 	u64 vdisktime = st->min_vdisktime;
 	struct cfq_group *cfqg;
 
-	if (st->active) {
-		cfqg = rb_entry_cfqg(st->active);
-		vdisktime = cfqg->vdisktime;
-	}
-
 	if (st->left) {
 		cfqg = rb_entry_cfqg(st->left);
 		vdisktime = min_vdisktime(vdisktime, cfqg->vdisktime);
@@ -894,9 +888,6 @@ cfq_group_service_tree_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
 {
 	struct cfq_rb_root *st = &cfqd->grp_service_tree;
 
-	if (st->active == &cfqg->rb_node)
-		st->active = NULL;
-
 	BUG_ON(cfqg->nr_cfqq < 1);
 	cfqg->nr_cfqq--;
 
@@ -1095,7 +1086,7 @@ static void cfq_put_cfqg(struct cfq_group *cfqg)
 	if (!atomic_dec_and_test(&cfqg->ref))
 		return;
 	for_each_cfqg_st(cfqg, i, j, st)
-		BUG_ON(!RB_EMPTY_ROOT(&st->rb) || st->active != NULL);
+		BUG_ON(!RB_EMPTY_ROOT(&st->rb));
 	kfree(cfqg);
 }
 
@@ -1687,9 +1678,6 @@ __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	if (cfqq == cfqd->active_queue)
 		cfqd->active_queue = NULL;
 
-	if (&cfqq->cfqg->rb_node == cfqd->grp_service_tree.active)
-		cfqd->grp_service_tree.active = NULL;
-
 	if (cfqd->active_cic) {
 		put_io_context(cfqd->active_cic->ioc);
 		cfqd->active_cic = NULL;
@@ -2199,7 +2187,6 @@ static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd)
 	if (RB_EMPTY_ROOT(&st->rb))
 		return NULL;
 	cfqg = cfq_rb_first_group(st);
-	st->active = &cfqg->rb_node;
 	update_min_vdisktime(st);
 	return cfqg;
 }
Signature · 1 line
1.6.5.2
maintainer
#2 lore raw
In reply to Gui Jianfeng
Quoted 2 lines · Gui Jianfeng

When a cfq group is running, it won't be dequeued from service tree, so there's no need to store the active one in st->active. Just gid rid of it.

Thanks Gui, applied.

Signature · 1 line
Jens Axboe