[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <b8deb6fa-8a09-c1af-278f-24e66afe367d@kernel.dk>
Date: Wed, 7 Dec 2022 10:22:09 -0700
From: Jens Axboe <axboe@...nel.dk>
To: Yu Kuai <yukuai1@...weicloud.com>, Ming Lei <ming.lei@...hat.com>
Cc: Gulam Mohamed <gulam.mohamed@...cle.com>,
linux-block@...r.kernel.org, philipp.reisner@...bit.com,
lars.ellenberg@...bit.com, christoph.boehmwalder@...bit.com,
minchan@...nel.org, ngupta@...are.org, senozhatsky@...omium.org,
colyli@...e.de, kent.overstreet@...il.com, agk@...hat.com,
snitzer@...nel.org, dm-devel@...hat.com, song@...nel.org,
dan.j.williams@...el.com, vishal.l.verma@...el.com,
dave.jiang@...el.com, ira.weiny@...el.com, junxiao.bi@...cle.com,
martin.petersen@...cle.com, kch@...dia.com,
drbd-dev@...ts.linbit.com, linux-kernel@...r.kernel.org,
linux-bcache@...r.kernel.org, linux-raid@...r.kernel.org,
nvdimm@...ts.linux.dev, konrad.wilk@...cle.com,
"yukuai (C)" <yukuai3@...wei.com>
Subject: Re: [RFC] block: Change the granularity of io ticks from ms to ns
On 12/7/22 6:09 AM, Yu Kuai wrote:
> Hi,
>
> 在 2022/12/07 11:15, Ming Lei 写道:
>> On Wed, Dec 07, 2022 at 10:19:08AM +0800, Yu Kuai wrote:
>>> Hi,
>>>
>>> 在 2022/12/07 2:15, Gulam Mohamed 写道:
>>>> Use ktime to change the granularity of IO accounting in block layer from
>>>> milli-seconds to nano-seconds to get the proper latency values for the
>>>> devices whose latency is in micro-seconds. After changing the granularity
>>>> to nano-seconds the iostat command, which was showing incorrect values for
>>>> %util, is now showing correct values.
>>>
>>> This patch didn't correct the counting of io_ticks, just make the
>>> error accounting from jiffies(ms) to ns. The problem that util can be
>>> smaller or larger still exist.
>>
>> Agree.
>>
>>>
>>> However, I think this change make sense consider that error margin is
>>> much smaller, and performance overhead should be minimum.
>>>
>>> Hi, Ming, how do you think?
>>
>> I remembered that ktime_get() has non-negligible overhead, is there any
>> test data(iops/cpu utilization) when running fio or t/io_uring on
>> null_blk with this patch?
>
> Yes, testing with null_blk is necessary, we don't want any performance
> regression.
null_blk is fine as a substitute, but I'd much rather run this on my
test bench with actual IO and devices.
> BTW, I thought it's fine because it's already used for tracking io
> latency.
Reading a nsec timestamp is a LOT more expensive than reading jiffies,
which is essentially free. If you look at the amount of work that's
gone into minimizing ktime_get() for the fast path in the IO stack,
then that's a testament to that.
So that's a very bad assumption, and definitely wrong.
--
Jens Axboe
Powered by blists - more mailing lists