All of lore.kernel.org
 help / color / mirror / Atom feed
From: Matt Mackall <mpm@selenic.com>
To: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Colin Leroy <colin@colino.net>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] let fat handle MS_SYNCHRONOUS flag
Date: Tue, 23 Nov 2004 21:35:52 -0800	[thread overview]
Message-ID: <20041124053552.GD2460@waste.org> (raw)
In-Reply-To: <87pt237se1.fsf@devron.myhome.or.jp>

On Wed, Nov 24, 2004 at 02:00:54PM +0900, OGAWA Hirofumi wrote:
> Matt Mackall <mpm@selenic.com> writes:
> 
> > On Wed, Nov 24, 2004 at 05:24:36AM +0900, OGAWA Hirofumi wrote:
> >> Colin Leroy <colin@colino.net> writes:
> >> 
> >> > It adds MS_SYNCHRONOUS support to FAT filesystem, so that less
> >> > filesystem breakage happen when disconnecting an USB key, for 
> >> > example. I'd like to have comments about it, because as it 
> >> > seems to work fine here, I'm not used to fs drivers and could
> >> > have made mistakes.
> >> 
> >> What cases should these patches guarantee that users can unplug the
> >> USB key?  And can we guarantee the same cases by improving autofs or
> >> the similar stuff?
> >
> > Well there can be no guarantees - there will always be a race between
> > flush and hot unplug. If we're careful with write ordering, we can
> > perhaps arrange to avoid the worst sorts of corruption, provided the
> > device does the right thing when it's in the middle of an IO.
> >
> > But generally I think this is a good idea as it shrinks the window.
> 
> Things which I want to say here - do we really need the bogus
> sync-mode?

I'm not sure why you say it's bogus. Ext2 for instance has long had a
mount option similar to this and it makes sense in volatile
environments. Having the flag in the superblock seems a sensible way
of doing it as well.
 
> Current fatfs is not keeping the consistency of data on the disk at
> all.  So, after all, the data on a disk is corrupting until all
> syscalls finish, right?

This is to protect against usage patters like mv a b, oh look, it's
done, unplug. Not lots of active readers/writers.

> If so, isn't this too slow? I doubt this is good solution for this
> problem (USB key unplugging)...

Perhaps. I think the behavior should be to honor the superblock flag
by default, overridable with -o sync|async at mount.
 
> Well, it seems good as start of sync-mode though.
> -- 
> OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>

-- 
Mathematics is the supreme nostalgia of our time.

  reply	other threads:[~2004-11-24  5:36 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-11-18 18:49 [PATCH] let fat handle MS_SYNCHRONOUS flag Colin Leroy
2004-11-18 18:51 ` [PATCH] let vfat " Colin Leroy
2004-11-23 19:32 ` [PATCH] let fat " Colin Leroy
2004-11-23 20:24 ` OGAWA Hirofumi
2004-11-24  3:20   ` Matt Mackall
2004-11-24  5:00     ` OGAWA Hirofumi
2004-11-24  5:35       ` Matt Mackall [this message]
2004-11-24  6:26         ` OGAWA Hirofumi
2004-11-24  6:40           ` Andrew Morton
2004-11-24  6:49             ` Matt Mackall
2004-11-24  6:57               ` Andrew Morton
2004-11-24 13:26             ` OGAWA Hirofumi
2004-11-24  5:41       ` Bernd Eckenfels
2004-11-24  7:30         ` Colin Leroy
2004-11-24  3:17 ` Matt Mackall
2004-11-24  7:34   ` Colin Leroy
2004-11-24  7:50     ` Matt Mackall
2004-11-24  8:40       ` Colin Leroy
2004-11-24 14:02 ` OGAWA Hirofumi
2004-11-24 14:25   ` Colin Leroy
2004-11-24 15:02   ` Colin Leroy
2004-11-24 18:10     ` OGAWA Hirofumi
2004-11-24 20:28       ` Colin Leroy
     [not found]         ` <87k6saunwl.fsf@devron.myhome.or.jp>
2004-11-25 12:47           ` Colin Leroy
     [not found]             ` <87act6ulc5.fsf@devron.myhome.or.jp>
2004-11-25 13:11               ` Colin Leroy
2004-12-29 19:33                 ` Robert Hardy
2004-12-30  3:02                   ` OGAWA Hirofumi

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=20041124053552.GD2460@waste.org \
    --to=mpm@selenic.com \
    --cc=colin@colino.net \
    --cc=hirofumi@mail.parknet.co.jp \
    --cc=linux-kernel@vger.kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.