From: YingHang Zhu <casualfisher@gmail.com>
To: Fengguang Wu <fengguang.wu@intel.com>
Cc: akpm@linux-foundation.org, linux-fsdevel@vger.kernel.org,
linux-kernel@vger.kernel.org, Dave Chinner <david@fromorbit.com>,
linux-mm@kvack.org, Ni zhan Chen <nizhan.chen@gmail.com>
Subject: Re: [PATCH] mm: readahead: remove redundant ra_pages in file_ra_state
Date: Fri, 26 Oct 2012 16:13:32 +0800 [thread overview]
Message-ID: <CAA9v8mF+D=bdn9QSNcMv81BSog1okkK=wCLHKGi6NBWoQzyC-w@mail.gmail.com> (raw)
In-Reply-To: <20121026080240.GA13662@localhost>
Hi Fengguang,
On Fri, Oct 26, 2012 at 4:02 PM, Fengguang Wu <fengguang.wu@intel.com> wrote:
> On Fri, Oct 26, 2012 at 03:47:19PM +0800, Ni zhan Chen wrote:
>> On 10/26/2012 03:36 PM, Fengguang Wu wrote:
>> >On Fri, Oct 26, 2012 at 03:19:57PM +0800, Ni zhan Chen wrote:
>> >>On 10/26/2012 03:09 PM, Fengguang Wu wrote:
>> >>>On Fri, Oct 26, 2012 at 03:03:12PM +0800, Ni zhan Chen wrote:
>> >>>>On 10/26/2012 02:58 PM, Fengguang Wu wrote:
>> >>>>>> static void shrink_readahead_size_eio(struct file *filp,
>> >>>>>> struct file_ra_state *ra)
>> >>>>>> {
>> >>>>>>- ra->ra_pages /= 4;
>> >>>>>>+ spin_lock(&filp->f_lock);
>> >>>>>>+ filp->f_mode |= FMODE_RANDOM;
>> >>>>>>+ spin_unlock(&filp->f_lock);
>> >>>>>>
>> >>>>>>As the example in comment above this function, the read maybe still
>> >>>>>>sequential, and it will waste IO bandwith if modify to FMODE_RANDOM
>> >>>>>>directly.
>> >>>>>Yes immediately disabling readahead may hurt IO performance, the
>> >>>>>original '/ 4' may perform better when there are only 1-3 IO errors
>> >>>>>encountered.
>> >>>>Hi Fengguang,
>> >>>>
>> >>>>Why the number should be 1-3?
>> >>>The original behavior is '/= 4' on each error.
>> >>>
>> >>>After 1 errors, readahead size will be shrinked by 1/4
>> >>>After 2 errors, readahead size will be shrinked by 1/16
>> >>>After 3 errors, readahead size will be shrinked by 1/64
>> >>>After 4 errors, readahead size will be effectively 0 (disabled)
>> >>But from function shrink_readahead_size_eio and its caller
>> >>filemap_fault I can't find the behavior you mentioned. How you
>> >>figure out it?
>> >It's this line in shrink_readahead_size_eio():
>> >
>> > ra->ra_pages /= 4;
>>
>> Yeah, I mean why the 4th readahead size will be 0(disabled)? What's
>> the original value of ra->ra_pages? How can guarantee the 4th shrink
>> readahead size can be 0?
>
> Ah OK, I'm talking about the typical case. The default readahead size
> is 128k, which will become 0 after / 256. The reasonable good ra size
> for hard disks is 1MB=256pages, which also becomes 1page after 4 errors.
How do you feel about my previous mail of error statistics, in fact I
prefer treating these files independently, so do some check for FMODE_RANDOM
before we change the readahead window to avoid trashing the read ahead window,
If the user applications call fadvise but the media turn out to be
error-prone, after one
read we know the situation and set the file in FMODE_RANDOM, this should solve
the issue Dave raised.
Do I miss something? Thanks in advance.
>
> Thanks,
> Fengguang
--
Thanks,
Ying Zhu
next prev parent reply other threads:[~2012-10-26 8:13 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-10-23 12:46 [PATCH] mm: readahead: remove redundant ra_pages in file_ra_state Ying Zhu
2012-10-23 13:21 ` Ni zhan Chen
[not found] ` <CAA9v8mGMa3SDD1OLTG_wdhCGx7K-0kvSV1+MRi9uCGTz6zZaLg@mail.gmail.com>
2012-10-23 13:41 ` YingHang Zhu
2012-10-24 1:02 ` Ni zhan Chen
2012-10-24 1:33 ` YingHang Zhu
2012-10-23 22:47 ` Dave Chinner
2012-10-23 23:53 ` YingHang Zhu
2012-10-24 20:19 ` Dave Chinner
2012-10-25 0:17 ` YingHang Zhu
2012-10-25 1:48 ` Ni zhan Chen
2012-10-25 1:50 ` Dave Chinner
2012-10-25 2:04 ` YingHang Zhu
2012-10-25 2:12 ` Ni zhan Chen
2012-10-25 2:31 ` YingHang Zhu
2012-10-25 2:58 ` Fengguang Wu
2012-10-25 3:12 ` YingHang Zhu
2012-10-26 0:25 ` Dave Chinner
2012-10-26 1:27 ` Fengguang Wu
2012-10-26 2:30 ` Ni zhan Chen
2012-10-26 3:28 ` YingHang Zhu
2012-10-26 3:51 ` Ni zhan Chen
2012-10-26 4:35 ` YingHang Zhu
2012-10-26 6:58 ` Fengguang Wu
2012-10-26 7:03 ` Ni zhan Chen
2012-10-26 7:09 ` Fengguang Wu
2012-10-26 7:19 ` Ni zhan Chen
2012-10-26 7:36 ` Fengguang Wu
2012-10-26 7:47 ` Ni zhan Chen
2012-10-26 8:02 ` Fengguang Wu
2012-10-26 8:08 ` Ni zhan Chen
2012-10-26 8:13 ` YingHang Zhu [this message]
2012-10-26 2:25 ` Ni zhan Chen
2012-10-26 3:38 ` YingHang Zhu
2012-10-26 3:55 ` Fengguang Wu
2012-10-26 5:00 ` YingHang Zhu
2012-10-25 2:38 ` Fengguang Wu
2012-10-25 3:08 ` YingHang Zhu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAA9v8mF+D=bdn9QSNcMv81BSog1okkK=wCLHKGi6NBWoQzyC-w@mail.gmail.com' \
--to=casualfisher@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=david@fromorbit.com \
--cc=fengguang.wu@intel.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=nizhan.chen@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).