[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <acc94cef-3d37-d1a3-21cc-3a514be3291e@kernel.org>
Date: Thu, 15 Dec 2022 10:21:41 +0800
From: Chao Yu <chao@...nel.org>
To: Jaegeuk Kim <jaegeuk@...nel.org>, Yangtao Li <frank.li@...o.com>
Cc: linux-f2fs-devel@...ts.sourceforge.net,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] f2fs: fix iostat parameter for discard
On 2022/12/14 3:11, Jaegeuk Kim wrote:
> On 12/13, Yangtao Li wrote:
>>> What do you think of extending this function to support io_counts?
>>>
>>> void f2fs_update_iostat(struct f2fs_sb_info *sbi, struct inode *inode,
>>> enum iostat_type type, unsigned long long io_bytes,
>>> unsigned long long io_counts)
>>
>> Support to have extra io_count.
>>
>> But I don't think there is any need to add additional parameters to f2fs_update_iostat.
>> IIUC, each call to f2fs_update_iostat means that the corresponding count increases by 1,
>> so only the internal processing of the function is required.
>>
>> BTW, let's type out the iocount of the additional record in the following way?
>>
>> time: 1670930162
>> [WRITE]
>> app buffered data: 4096(1)
>
> How about giving in another columns with additional stats like avg. len/call or max. len?
Maybe call is better? w/ it we can calculate avg. len/call.
Thanks,
>
> app buffered data: 4096 1
>
>>
>> Thx,
>> Yangtao
Powered by blists - more mailing lists