Re: [RFC] blk-mq: clean up the hctx restart

[Bart Van Assche]

On Fri, 2018-08-03 at 00:58 +-0800, Ming Lei wrote:
+AD4- And about the situations you mentioned, no any special as normal cases
+AD4- or thousands of LUNs. Just a batch of queues are waken up from one
+AD4- single wait queue(sbq+AF8-wait+AF8-state), and inside each wait queue, queues
+AD4- are handled actually in FIFO order.
+AD4- 
+AD4- 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 treated
equally. Unless someone can set up a test that demonstrates that this is still
the case for commit 97889f9ac24f (+ACI-blk-mq: remove synchronize+AF8-rcu() from
blk+AF8-mq+AF8-del+AF8-queue+AF8-tag+AF8-set()+ACI-), I will assume that that commit breaks fairness.

Bart.