block: add a switch for precise iostat accounting

hulk inclusion
category: bugfix
bugzilla: 39265, https://gitee.com/openeuler/kernel/issues/I4WC06


CVE: NA

-----------------------------------------------

When the inflight IOs are slow and no new IOs are issued, we expect
iostat could manifest the IO hang problem. However after
commit 9c6dea45 ("block: delete part_round_stats and switch to less
precise counting"), io_tick and time_in_queue will not be updated until
the end of IO, and the avgqu-sz and %util columns of iostat will be zero.

To fix it, we could fallback to the implementation before commit
9c6dea45, but it may cause performance regression on NVMe device
or bio-based device (due to overhead of inflight calculation),
so add a switch to control whether or not to use precise iostat
accounting. It can be enabled by adding "precise_iostat=1" in kernel
boot cmdline. When precise accouting is enabled, io_tick and time_in_queue
will be updated when accessing /proc/diskstats and
/sys/block/sdX/sdXN/stat.

Fixes: 9c6dea45 ("block: delete part_round_stats and switch to less precise counting")
Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>