blkcg: implement per-queue policy activation

All blkcg policies were assumed to be enabled on all request_queues.
Due to various implementation obstacles, during the recent blkcg core
updates, this was temporarily implemented as shooting down all !root
blkgs on elevator switch and policy [de]registration combined with
half-broken in-place root blkg updates.  In addition to being buggy
and racy, this meant losing all blkcg configurations across those
events.

Now that blkcg is cleaned up enough, this patch replaces the temporary
implementation with proper per-queue policy activation.  Each blkcg
policy should call the new blkcg_[de]activate_policy() to enable and
disable the policy on a specific queue.  blkcg_activate_policy()
allocates and installs policy data for the policy for all existing
blkgs.  blkcg_deactivate_policy() does the reverse.  If a policy is
not enabled for a given queue, blkg printing / config functions skip
the respective blkg for the queue.

blkcg_activate_policy() also takes care of root blkg creation, and
cfq_init_queue() and blk_throtl_init() are updated accordingly.

This replaces blkcg_bypass_{start|end}() and update_root_blkg_pd()
unnecessary.  Dropped.

v2: cfq_init_queue() was returning uninitialized @ret on root_group
    alloc failure if !CONFIG_CFQ_GROUP_IOSCHED.  Fixed.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

1 files changed, 16 insertions, 21 deletions

diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 86440e04f3ee..0203652e1f34 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -1406,8 +1406,7 @@ static int cfqg_set_weight_device(struct cgroup *cgrp, struct cftype *cft,
 
         ret = -EINVAL;
         cfqg = blkg_to_cfqg(ctx.blkg);
-        if (cfqg && (!ctx.v || (ctx.v >= CFQ_WEIGHT_MIN &&
-                                ctx.v <= CFQ_WEIGHT_MAX))) {
+        if (!ctx.v || (ctx.v >= CFQ_WEIGHT_MIN && ctx.v <= CFQ_WEIGHT_MAX)) {
                 cfqg->dev_weight = ctx.v;
                 cfqg->new_weight = cfqg->dev_weight ?: blkcg->cfq_weight;
                 ret = 0;
@@ -3938,7 +3937,7 @@ static void cfq_exit_queue(struct elevator_queue *e)
 #ifndef CONFIG_CFQ_GROUP_IOSCHED
         kfree(cfqd->root_group);
 #endif
-        update_root_blkg_pd(q, &blkio_policy_cfq);
+        blkcg_deactivate_policy(q, &blkio_policy_cfq);
         kfree(cfqd);
 }
 
@@ -3946,7 +3945,7 @@ static int cfq_init_queue(struct request_queue *q)
 {
         struct cfq_data *cfqd;
         struct blkio_group *blkg __maybe_unused;
-        int i;
+        int i, ret;
 
         cfqd = kmalloc_node(sizeof(*cfqd), GFP_KERNEL | __GFP_ZERO, q->node);
         if (!cfqd)
@@ -3960,28 +3959,20 @@ static int cfq_init_queue(struct request_queue *q)
 
         /* Init root group and prefer root group over other groups by default */
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-        rcu_read_lock();
-        spin_lock_irq(q->queue_lock);
-
-        blkg = blkg_lookup_create(&blkio_root_cgroup, q, true);
-        if (!IS_ERR(blkg)) {
-                q->root_blkg = blkg;
-                cfqd->root_group = blkg_to_cfqg(blkg);
-        }
+        ret = blkcg_activate_policy(q, &blkio_policy_cfq);
+        if (ret)
+                goto out_free;
 
-        spin_unlock_irq(q->queue_lock);
-        rcu_read_unlock();
+        cfqd->root_group = blkg_to_cfqg(q->root_blkg);
 #else
+        ret = -ENOMEM;
         cfqd->root_group = kzalloc_node(sizeof(*cfqd->root_group),
                                         GFP_KERNEL, cfqd->queue->node);
-        if (cfqd->root_group)
-                cfq_init_cfqg_base(cfqd->root_group);
-#endif
-        if (!cfqd->root_group) {
-                kfree(cfqd);
-                return -ENOMEM;
-        }
+        if (!cfqd->root_group)
+                goto out_free;
 
+        cfq_init_cfqg_base(cfqd->root_group);
+#endif
         cfqd->root_group->weight = 2 * CFQ_WEIGHT_DEFAULT;
 
         /*
@@ -4031,6 +4022,10 @@ static int cfq_init_queue(struct request_queue *q)
          */
         cfqd->last_delayed_sync = jiffies - HZ;
         return 0;
+
+out_free:
+        kfree(cfqd);
+        return ret;
 }
 
 /*