On Fri, 2018-08-03 at 00:58 +0800, Ming Lei wrote:

And about the situations you mentioned, no any special as normal case=

s

or thousands of LUNs. Just a batch of queues are waken up from one
single wait queue(sbq_wait_state), and inside each wait queue=

, queues

are handled actually in FIFO order.
=20
Or what is your expected ideal behaviour about fairness?

Hello Ming,

What I expect is if the number of LUNs is really large that all LUNs are tr=
eated
equally. Unless someone can set up a test that demonstrates that this is st=
ill
the case for commit 97889f9ac24f ("blk-mq: remove synchronize_rcu()=
 from
blk_mq_del_queue_tag_set()"), I will assume that th=
at commit breaks fairness.

Bart.