All of lore.kernel.org
 help / color / mirror / Atom feed
From: Daniel Phillips <phillips@arcor.de>
To: David Howells <dhowells@redhat.com>
Cc: Andrew Morton <akpm@osdl.org>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	hugh@veritas.com
Subject: Re: [RFC][patch 0/2] mm: remove PageReserved
Date: Tue, 16 Aug 2005 11:53:58 +1000	[thread overview]
Message-ID: <200508161153.59476.phillips@arcor.de> (raw)
In-Reply-To: <8985.1124111717@warthog.cambridge.redhat.com>

On Monday 15 August 2005 23:15, David Howells wrote:
> I want to know when a page is going to be modified so that I
> can predict the state of the cache as much as possible. I don't want
> userspace processes corrupting the cache in unrecorded ways.

There are two cases:

  1) Metadata.  If anybody is doing racy writes to metadata pages, it is
     your filesystem, and you have a bug.

  2) Data.  In Linux practice and Posix, racy writes to files have
     undefined semantics, including the possibility that data may end up
     interleaved on a disk block.

You seem to be trying to define (2) as "corruption" and setting out to prevent 
it.  But it is not the responsibility of a filesystem to prevent this, it is 
the responsibility of the application.

Could you please explain why it is not ok to end up with a half-written page 
in your cache, if the client was in fact halfway through writing it when it 
crashed?

Regards,

Daniel

WARNING: multiple messages have this Message-ID (diff)
From: Daniel Phillips <phillips@arcor.de>
To: David Howells <dhowells@redhat.com>
Cc: Andrew Morton <akpm@osdl.org>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	hugh@veritas.com
Subject: Re: [RFC][patch 0/2] mm: remove PageReserved
Date: Tue, 16 Aug 2005 11:53:58 +1000	[thread overview]
Message-ID: <200508161153.59476.phillips@arcor.de> (raw)
In-Reply-To: <8985.1124111717@warthog.cambridge.redhat.com>

On Monday 15 August 2005 23:15, David Howells wrote:
> I want to know when a page is going to be modified so that I
> can predict the state of the cache as much as possible. I don't want
> userspace processes corrupting the cache in unrecorded ways.

There are two cases:

  1) Metadata.  If anybody is doing racy writes to metadata pages, it is
     your filesystem, and you have a bug.

  2) Data.  In Linux practice and Posix, racy writes to files have
     undefined semantics, including the possibility that data may end up
     interleaved on a disk block.

You seem to be trying to define (2) as "corruption" and setting out to prevent 
it.  But it is not the responsibility of a filesystem to prevent this, it is 
the responsibility of the application.

Could you please explain why it is not ok to end up with a half-written page 
in your cache, if the client was in fact halfway through writing it when it 
crashed?

Regards,

Daniel
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2005-08-16  1:53 UTC|newest]

Thread overview: 180+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-08-07  3:28 [RFC][patch 0/2] mm: remove PageReserved Nick Piggin
2005-08-07  3:28 ` Nick Piggin
2005-08-07  3:29 ` [patch 1/2] mm: remap ZERO_PAGE mappings Nick Piggin
2005-08-07  3:30   ` [patch 2/2] mm: core remove PageReserved Nick Piggin
2005-08-08 21:09 ` [RFC][patch 0/2] mm: " Daniel Phillips
2005-08-08 21:09   ` Daniel Phillips
2005-08-08 21:24   ` Daniel Phillips
2005-08-08 21:24     ` Daniel Phillips
2005-08-08 21:54     ` Andrew Morton
2005-08-08 21:54       ` Andrew Morton
2005-08-09 23:23       ` [RFC][PATCH] Rename PageChecked as PageMiscFS Daniel Phillips
2005-08-09 23:23         ` Daniel Phillips
2005-08-10  7:48         ` Hugh Dickins
2005-08-10  7:48           ` Hugh Dickins
2005-08-10  8:06           ` Daniel Phillips
2005-08-10  8:06             ` Daniel Phillips
2005-08-10 13:13       ` [RFC][patch 0/2] mm: remove PageReserved David Howells
2005-08-10 13:13         ` David Howells
2005-08-10 13:34         ` Daniel Phillips
2005-08-10 13:34           ` Daniel Phillips
2005-08-10 14:27           ` David Howells
2005-08-10 14:27             ` David Howells
2005-08-10 23:19             ` Daniel Phillips
2005-08-10 23:19               ` Daniel Phillips
2005-08-11 10:49               ` David Howells
2005-08-11 10:49                 ` David Howells
2005-08-12 19:34                 ` Daniel Phillips
2005-08-12 19:34                   ` Daniel Phillips
2005-08-15 13:15                   ` David Howells
2005-08-15 13:15                     ` David Howells
2005-08-16  1:53                     ` Daniel Phillips [this message]
2005-08-16  1:53                       ` Daniel Phillips
2005-08-16 10:28                       ` David Howells
2005-08-16 10:28                         ` David Howells
2005-08-10 22:12       ` [RFC][PATCH] Rename PageChecked as PageMiscFS Daniel Phillips
2005-08-10 22:12         ` Daniel Phillips
2005-08-10 22:23         ` Daniel Phillips
2005-08-10 22:23           ` Daniel Phillips
2005-08-10 22:34           ` Trond Myklebust
2005-08-10 22:34             ` Trond Myklebust
2005-08-10 22:57             ` Daniel Phillips
2005-08-10 22:57               ` Daniel Phillips
2005-08-10 23:23               ` Trond Myklebust
2005-08-10 23:23                 ` Trond Myklebust
2005-08-11  9:42                 ` David Howells
2005-08-11  9:42                   ` David Howells
2005-08-10 23:42           ` Adrian Bunk
2005-08-10 23:42             ` Adrian Bunk
2005-08-11  9:46             ` David Howells
2005-08-11  9:46               ` David Howells
2005-08-12  2:34               ` Daniel Phillips
2005-08-12  2:34                 ` Daniel Phillips
2005-08-12 12:32                 ` David Howells
2005-08-11  9:31           ` David Howells
2005-08-11  9:31             ` David Howells
2005-08-11  9:26         ` David Howells
2005-08-11  9:26           ` David Howells
2005-08-12  3:29           ` Daniel Phillips
2005-08-12  3:29             ` Daniel Phillips
2005-08-12 12:41             ` David Howells
2005-08-12 12:41               ` David Howells
2005-08-12 13:28               ` Hugh Dickins
2005-08-12 13:28                 ` Hugh Dickins
2005-08-16 13:59               ` Pavel Machek
2005-08-16 13:59                 ` Pavel Machek
2005-08-18 14:33                 ` David Howells
2005-08-18 14:33                   ` David Howells
2005-08-18 22:27                   ` Pavel Machek
2005-08-18 22:27                     ` Pavel Machek
2005-08-19 10:04                     ` David Howells
2005-08-19 10:04                       ` David Howells
2005-08-19 16:31                       ` Daniel Phillips
2005-08-19 16:31                         ` Daniel Phillips
2005-08-20 10:45                         ` David Howells
2005-08-20 10:45                           ` David Howells
2005-08-20 20:21                           ` Daniel Phillips
2005-08-20 20:21                             ` Daniel Phillips
2005-08-09  0:15   ` [RFC][patch 0/2] mm: remove PageReserved Nick Piggin
2005-08-09  0:15     ` Nick Piggin
2005-08-09  8:51     ` Benjamin Herrenschmidt
2005-08-09  8:51       ` Benjamin Herrenschmidt
2005-08-09  9:49       ` Nick Piggin
2005-08-09  9:49         ` Nick Piggin
2005-08-09 19:19         ` Daniel Phillips
2005-08-09 19:19           ` Daniel Phillips
2005-08-09 19:22         ` Daniel Phillips
2005-08-09 19:22           ` Daniel Phillips
2005-08-10 21:50           ` Pavel Machek
2005-08-10 21:50             ` Pavel Machek
2005-08-10 21:56             ` Martin J. Bligh
2005-08-10 21:56               ` Martin J. Bligh
2005-08-11 10:36               ` Rafael J. Wysocki
2005-08-11 10:36                 ` Rafael J. Wysocki
2005-08-12 19:56                 ` Daniel Phillips
2005-08-12 19:56                   ` Daniel Phillips
2005-08-12 22:20                   ` Rafael J. Wysocki
2005-08-12 22:20                     ` Rafael J. Wysocki
2005-08-12 23:04                     ` Daniel Phillips
2005-08-12 23:04                       ` Daniel Phillips
2005-08-13  7:06                       ` Rafael J. Wysocki
2005-08-13  7:06                         ` Rafael J. Wysocki
2005-08-11 10:26             ` Rafael J. Wysocki
2005-08-11 10:26               ` Rafael J. Wysocki
2005-08-09 11:25       ` Hugh Dickins
2005-08-09 11:25         ` Hugh Dickins
2005-08-09 14:31         ` Benjamin Herrenschmidt
2005-08-09 14:31           ` Benjamin Herrenschmidt
2005-08-09 14:50           ` Hugh Dickins
2005-08-09 14:50             ` Hugh Dickins
2005-08-09 14:49             ` Benjamin Herrenschmidt
2005-08-09 14:49               ` Benjamin Herrenschmidt
2005-08-09 15:36               ` Hugh Dickins
2005-08-09 15:36                 ` Hugh Dickins
2005-08-09 21:27                 ` Daniel Phillips
2005-08-09 21:27                   ` Daniel Phillips
2005-08-09 19:14     ` Daniel Phillips
2005-08-09 19:14       ` Daniel Phillips
2005-08-09 20:17       ` Hugh Dickins
2005-08-09 20:17         ` Hugh Dickins
2005-08-09 20:52         ` Daniel Phillips
2005-08-09 20:52           ` Daniel Phillips
2005-08-09  4:39   ` Nigel Cunningham
2005-08-09  4:39     ` Nigel Cunningham
2005-08-09  4:59     ` Nick Piggin
2005-08-09  4:59       ` Nick Piggin
2005-08-09  5:11       ` Nigel Cunningham
2005-08-09  5:11         ` Nigel Cunningham
2005-08-09  5:20         ` Nick Piggin
2005-08-09  5:20           ` Nick Piggin
2005-08-09  5:30           ` Nigel Cunningham
2005-08-09  5:30             ` Nigel Cunningham
2005-08-09  7:08       ` Russell King
2005-08-09  7:08         ` Russell King
2005-08-09  8:38         ` Arjan van de Ven
2005-08-09  8:38           ` Arjan van de Ven
2005-08-09  9:31           ` Nick Piggin
2005-08-09  9:31             ` Nick Piggin
2005-08-09  9:49             ` Arjan van de Ven
2005-08-09  9:49               ` Arjan van de Ven
2005-08-09  9:57               ` Nick Piggin
2005-08-09  9:57                 ` Nick Piggin
2005-08-09 10:24             ` Rafael J. Wysocki
2005-08-09 10:24               ` Rafael J. Wysocki
2005-08-09  8:53         ` Benjamin Herrenschmidt
2005-08-09  8:53           ` Benjamin Herrenschmidt
2005-08-09  9:15         ` Hugh Dickins
2005-08-09  9:15           ` Hugh Dickins
2005-08-09 10:27           ` Nick Piggin
2005-08-09 10:27             ` Nick Piggin
2005-08-09 11:15             ` Hugh Dickins
2005-08-09 11:15               ` Hugh Dickins
2005-08-09 13:15               ` Nick Piggin
2005-08-09 13:15                 ` Nick Piggin
2005-08-09 13:26                 ` Arjan van de Ven
2005-08-09 13:26                   ` Arjan van de Ven
2005-08-09 14:28               ` Benjamin Herrenschmidt
2005-08-09 14:28                 ` Benjamin Herrenschmidt
2005-08-09 14:47                 ` Hugh Dickins
2005-08-09 14:47                   ` Hugh Dickins
2005-08-09 19:49           ` Roman Zippel
2005-08-09 19:49             ` Roman Zippel
2005-08-09  9:29         ` Nick Piggin
2005-08-09  9:29           ` Nick Piggin
2005-08-09 19:40           ` Russell King
2005-08-09 19:40             ` Russell King
2005-08-09 14:38         ` Martin J. Bligh
2005-08-09 14:38           ` Martin J. Bligh
2005-08-09 19:41           ` Russell King
2005-08-09 19:41             ` Russell King
2005-08-09 20:51             ` Linus Torvalds
2005-08-09 20:51               ` Linus Torvalds
2005-08-09 21:16             ` Martin J. Bligh
2005-08-09 21:16               ` Martin J. Bligh
2005-08-09 21:51               ` Martin J. Bligh
2005-08-09 21:51                 ` Martin J. Bligh
2005-08-10  9:27             ` Benjamin Herrenschmidt
2005-08-10  9:27               ` Benjamin Herrenschmidt
2005-08-11  9:09               ` Nick Piggin
2005-08-11  9:09                 ` Nick Piggin
2005-08-09 22:14 ` Daniel Phillips

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=200508161153.59476.phillips@arcor.de \
    --to=phillips@arcor.de \
    --cc=akpm@osdl.org \
    --cc=dhowells@redhat.com \
    --cc=hugh@veritas.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.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.