From: Nick Piggin <nickpiggin@yahoo.com.au>
To: Andrew Morton <akpm@osdl.org>
Cc: linux@horizon.com, linux-kernel@vger.kernel.org, sct@redhat.com
Subject: Re: msync() behaviour broken for MS_ASYNC, revert patch?
Date: Fri, 10 Feb 2006 15:30:16 +1100 [thread overview]
Message-ID: <43EC16D8.8030300@yahoo.com.au> (raw)
In-Reply-To: <20060209201333.62db0e24.akpm@osdl.org>
Andrew Morton wrote:
> Nick Piggin <nickpiggin@yahoo.com.au> wrote:
>
>> I don't think anyone would use MS_ASYNC for anything other than
>> performance improvement, so it is not like we need super well
>> defined behaviour... the earlier it will start IO AFAIKS the better.
>
>
> Well, no. Consider a continuously-running application which modifies its
> data store via MAP_SHARED+msync(MS_ASYNC). If the msync() immediately
> started I/O, the disk would be seeking all over the place all the time. The
> queue merging and timer-based unplugging would help here, but it won't be
> as good as a big, infrequent ascending-file-offset pdflush pass.
>
Sure you can shoot yourself in the foot.
"msync flushes changes made to the in-core copy of a file that
was mapped into memory using mmap(2) back to disk. "
We usually don't cater to foot shooters at the expense of valid users.
AFAIKS, basically the only valid use for MS_ASYNC is for the app to tell
the kernel that it isn't going to write here for a long time, so writeback
may as well be scheduled; or to pipeline other work with an upcoming data
integrity point which will need to be guaranteed by a second call to MS_SYNC.
> Secondly, consider the behaviour of the above application if it is modifying
> the same page relatively frequently (quite likely). If MS_ASYNC starts I/O
> immediately, that page will get written 10, 100 or 1000 times per second.
> If MS_ASYNC leaves it to pdflush, that page gets written once per 30
> seconds, so we do far much less I/O.
>
> We just don't know. It's better to leave it up to the application designer
> rather than lumping too many operations into the one syscall.
Well it remains the same conceptual operation (asynchronously "schedule"
dirty pages for writeout). However it simply becomes more useful to start
the writeout immediately, given that's the (pretty explicit) hint that is
given to us.
--
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com
next prev parent reply other threads:[~2006-02-10 4:30 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-02-09 7:18 msync() behaviour broken for MS_ASYNC, revert patch? linux
2006-02-09 8:18 ` Andrew Morton
2006-02-09 8:35 ` Nick Piggin
2006-02-09 8:42 ` Andrew Morton
2006-02-09 12:38 ` Nick Piggin
2006-02-09 12:39 ` Nick Piggin
2006-02-09 17:48 ` Andrew Morton
2006-02-10 3:36 ` Nick Piggin
2006-02-10 3:50 ` Andrew Morton
2006-02-10 3:57 ` Nick Piggin
2006-02-10 4:13 ` Andrew Morton
2006-02-10 4:30 ` Nick Piggin [this message]
2006-02-10 4:43 ` Andrew Morton
2006-02-10 4:52 ` Nick Piggin
2006-02-10 5:13 ` Andrew Morton
2006-02-10 5:29 ` Nick Piggin
2006-02-10 5:50 ` Andrew Morton
2006-02-10 6:03 ` Nick Piggin
2006-02-10 6:13 ` Andrew Morton
2006-02-10 6:31 ` Nick Piggin
2006-02-10 6:46 ` Andrew Morton
2006-02-10 6:57 ` Nick Piggin
2006-02-10 7:14 ` Andrew Morton
2006-02-10 12:41 ` Nick Piggin
2006-02-10 16:19 ` Linus Torvalds
2006-02-10 17:00 ` Nick Piggin
2006-02-10 17:12 ` Linus Torvalds
2006-02-10 17:35 ` Linus Torvalds
2006-02-10 17:59 ` Nick Piggin
2006-02-10 18:55 ` Linus Torvalds
2006-02-10 19:29 ` Nick Piggin
2006-02-10 19:44 ` Linus Torvalds
2006-02-10 19:52 ` Nick Piggin
2006-02-10 20:03 ` Linus Torvalds
2006-02-11 5:49 ` Nick Piggin
2006-02-10 16:05 ` Linus Torvalds
2006-02-10 16:37 ` Nick Piggin
2006-02-10 17:03 ` Linus Torvalds
2006-02-10 17:37 ` Nick Piggin
2006-02-10 18:01 ` Linus Torvalds
2006-02-10 18:38 ` Nick Piggin
2006-02-10 19:05 ` Linus Torvalds
2006-02-10 19:34 ` Oliver Neukum
2006-02-10 19:59 ` Linus Torvalds
2006-02-10 20:11 ` Andrew Morton
2006-02-10 21:15 ` Linus Torvalds
2006-02-10 21:28 ` Andrew Morton
2006-02-10 20:03 ` Nick Piggin
2006-02-10 21:10 ` Linus Torvalds
2006-02-10 21:55 ` Trond Myklebust
2006-02-10 22:46 ` Linus Torvalds
2006-02-10 23:02 ` Trond Myklebust
2006-02-10 23:15 ` Linus Torvalds
2006-02-11 19:07 ` Trond Myklebust
2006-02-10 17:29 ` linux
2006-02-10 17:42 ` Linus Torvalds
2006-02-10 18:57 ` Nick Piggin
2006-02-10 8:00 ` linux
2006-02-10 13:18 ` Nick Piggin
2006-02-10 7:15 ` linux
2006-02-10 7:28 ` Andrew Morton
2006-02-09 11:18 ` linux
-- strict thread matches above, loose matches on Subject: below --
2004-03-31 22:16 Stephen C. Tweedie
2004-03-31 22:37 ` Linus Torvalds
2004-03-31 23:41 ` Stephen C. Tweedie
2004-04-01 0:08 ` Linus Torvalds
2004-04-01 0:30 ` Andrew Morton
2004-04-01 15:40 ` Stephen C. Tweedie
2004-04-01 16:02 ` Linus Torvalds
2004-04-01 16:33 ` Stephen C. Tweedie
2004-04-01 16:19 ` Jamie Lokier
2004-04-01 16:57 ` Stephen C. Tweedie
2004-04-01 18:51 ` Andrew Morton
2004-03-31 22:53 ` Andrew Morton
2004-03-31 23:20 ` Stephen C. Tweedie
2004-04-16 22:35 ` Jamie Lokier
2004-04-19 21:54 ` Stephen C. Tweedie
2004-04-21 2:10 ` Jamie Lokier
2004-04-21 9:52 ` Stephen C. Tweedie
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=43EC16D8.8030300@yahoo.com.au \
--to=nickpiggin@yahoo.com.au \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@horizon.com \
--cc=sct@redhat.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