block: fix inaccurate io_ticks by set 'precise_iostat'
hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I5X6RT CVE: NA -------------------------------- After introducing commit 5b18b5a7 ("block: delete part_round_stats and switch to less precise counting"), '%util' accounted by iostat will be over reality data. In fact, the device is quite idle, but iostat may show '%util' as a big number (e.g. 50%). It can produce by fio: fio --name=1 --direct=1 --bs=4k --rw=read --filename=/dev/sda \ --thinktime=4ms --runtime=180 We fix this by using a switch(precise_iostat=1) to control whether or not acconut ioticks precisely. fixes: 5b18b5a7 ("block: delete part_round_stats and switch to less precise counting") Signed-off-by: NZhang Wensheng <zhangwensheng5@huawei.com> Signed-off-by: NYu Kuai <yukuai3@huawei.com> Reviewed-by: NJason Yan <yanaijie@huawei.com> Signed-off-by: NZheng Zengkai <zhengzengkai@huawei.com>
Showing
想要评论请 注册 或 登录