Hello, Bart. On Tue, Jan 09, 2018 at 04:12:40PM +0000, Bart Van Assche wrote: > I'm concerned about the additional CPU cycles needed for the new > blk_mq_map_queue() > call, although I know this call is cheap. Would the timeout code really get > that
So, if that is really a concern, let's cache that mapping instead of changing synchronization rules for that. > much more complicated if the hctx_lock() and hctx_unlock() calls would be > changed > into rcu_read_lock() and rcu_read_unlock() calls? Would it be sufficient if > "if (has_rcu) synchronize_rcu();" would be changed into "synchronize_rcu();" > in > blk_mq_timeout_work()? Code-wise, it won't be too much extra code but I think diverging the sync methods between issue and completion paths is more fragile and likely to invite confusions and mistakes in the future. We have the normal path (issue&completion) synchronizing against the exception path (timeout). I think it's best to keep the sync constructs aligned with that conceptual picture. Thanks. -- tejun
