On 2/26/19 6:57 PM, Jens Axboe wrote:

Speaking of this, I took a quick look at why we've now regressed a lot
on IOPS perf with the multipage work. It looks like it's all related to
the (much) fatter setup around iteration, which is related to this very
topic too.

Basically setup of things like bio_for_each_bvec() and indexing through
nth_page() is MUCH slower than before.

We need to do something about this, it's like tossing out months of
optimizations.

See attached quick profile. This is doing 4k IOS, btw, so just single
segment requests. Yet we spend most of the time in blk_rq_map_sg(), and
blk_queue_split() is not far behind. Between just the two of them,
that's more than 8% of the total CPU utilization. As this test case is
running on just one CPU for all of it, that's a LOT of wasted time.

-- 
Jens Axboe