qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Alex Bligh <alex@alex.org.uk>
To: Paolo Bonzini <pbonzini@redhat.com>,
	George Dunlap <george.dunlap@eu.citrix.com>
Cc: Ian Campbell <Ian.Campbell@citrix.com>,
	Stefano Stabellini <Stefano.Stabellini@eu.citrix.com>,
	Alex Bligh <alex@alex.org.uk>,
	Ian Jackson <Ian.Jackson@eu.citrix.com>,
	qemu-devel@nongnu.org, xen-devel <xen-devel@lists.xen.org>,
	Anthony Liguori <anthony@codemonkey.ws>
Subject: Re: [Qemu-devel] [Xen-devel] [PATCHv3] QEMU(upstream): Disable xen's use of O_DIRECT by default as it results in crashes.
Date: Tue, 19 Mar 2013 11:44:22 +0000	[thread overview]
Message-ID: <45EC1ED73B198669B314A507@nimrod.local> (raw)
In-Reply-To: <5148489C.70808@redhat.com>



--On 19 March 2013 12:14:36 +0100 Paolo Bonzini <pbonzini@redhat.com> wrote:

> Unfortunately no.  The problem is in the loading side's kernel, on which
> you do not have any control.  If the loading side doesn't use O_DIRECT,
> any attempt to invalidate the metadata in userspace or on the source is
> futile, because there is no way to invalidate the page cache's copy of
> that metadata.

If you closed the file and reopened it, then on NFS this would have this
effect I believe, as it needs to check whether other clients have modified
the file to give open/close coherency.

However, we are rather assuming the file isn't even open at the point it
is closed by the sender, which is what George is investigating.

If this isn't true, we have a problem anyway with (e.g.) emulated
devices which don't use O_DIRECT anyway. And I had thought (I may be
wrong) using O_DIRECT does not guarantee no read caching with NFS;
O_DIRECT merely guarantees the page cache is not used under Linux
and isn't defined under POSIX:
  http://www.spinics.net/lists/linux-nfs/msg17472.html

If it were just a write caching issue, we could use O_DSYNC instead of
O_DIRECT, which would at least ensure the copy from userspace.

-- 
Alex Bligh

  parent reply	other threads:[~2013-03-19 11:44 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-03-18 12:18 [Qemu-devel] [PATCHv3] QEMU(upstream): Disable xen's use of O_DIRECT by default as it results in crashes Alex Bligh
2013-03-18 13:03 ` Stefan Hajnoczi
2013-03-18 13:19   ` Alex Bligh
2013-03-18 13:32 ` Paolo Bonzini
2013-03-18 13:49   ` Alex Bligh
2013-03-18 14:05     ` Paolo Bonzini
2013-03-18 14:30       ` Alex Bligh
2013-03-18 14:49         ` Paolo Bonzini
2013-03-18 15:40           ` Alex Bligh
2013-03-18 16:19             ` Paolo Bonzini
2013-03-18 16:53               ` Alex Bligh
2013-03-18 17:38                 ` George Dunlap
2013-03-18 17:47                   ` Alex Bligh
2013-03-18 18:00                   ` Paolo Bonzini
2013-03-19 10:06                     ` [Qemu-devel] [Xen-devel] " George Dunlap
2013-03-19 10:43                       ` Paolo Bonzini
2013-03-19 10:51                         ` George Dunlap
2013-03-19 11:14                           ` Paolo Bonzini
2013-03-19 11:21                             ` George Dunlap
2013-03-19 15:12                               ` George Dunlap
2013-03-19 15:29                                 ` George Dunlap
2013-03-19 19:15                                   ` Alex Bligh
2013-03-20 10:24                                     ` Stefano Stabellini
2013-03-20 10:37                                       ` George Dunlap
2013-03-20 11:08                                         ` Paolo Bonzini
2013-03-20 11:20                                           ` Alex Bligh
2013-03-20 11:57                                       ` David Scott
2013-03-19 11:44                             ` Alex Bligh [this message]
2013-03-19 11:49                               ` Paolo Bonzini
2013-03-19 15:13                         ` Stefano Stabellini
2013-03-19 16:53                           ` Paolo Bonzini
2013-03-19 17:03                             ` Stefano Stabellini
2013-03-20  8:33                           ` Alex Bligh
2013-03-20  9:26                             ` Paolo Bonzini
2013-03-29 17:19                               ` Stefano Stabellini
2013-03-31 19:53                                 ` Alex Bligh
2013-04-01 15:32                                   ` [Qemu-devel] [PATCH] [RFC] Xen PV backend: Move call to bdrv_new from blk_init to blk_connect Alex Bligh
2013-04-01 15:44                                     ` Stefano Stabellini
2013-04-01 20:56                                       ` Alex Bligh
2013-04-02 11:08                                         ` Stefano Stabellini
2013-04-05 10:31                                           ` [Qemu-devel] [PATCHv2 1/2] " Alex Bligh
2013-04-05 10:31                                             ` [Qemu-devel] [PATCHv2 2/2] Xen PV backend: Disable use of O_DIRECT by default as it results in crashes Alex Bligh
2013-04-05 14:22                                             ` [Qemu-devel] [PATCHv2 1/2] Xen PV backend: Move call to bdrv_new from blk_init to blk_connect Stefano Stabellini
2013-04-05 15:42                                               ` [Qemu-devel] [PATCHv3 " Alex Bligh
2013-04-05 15:42                                                 ` [Qemu-devel] [PATCHv3 2/2] Xen PV backend: Disable use of O_DIRECT by default as it results in crashes Alex Bligh
2013-04-05 15:43                                               ` [Qemu-devel] [PATCHv2 1/2] Xen PV backend: Move call to bdrv_new from blk_init to blk_connect Alex Bligh
2013-04-05 10:32                                           ` [Qemu-devel] [PATCHv2 1/2] Xen PV backend (for qemu-upstream-4.2-testing): " Alex Bligh
2013-04-05 10:32                                             ` [Qemu-devel] [PATCHv2 2/2] Xen PV backend (for qemu-upstream-4.2-testing): Disable use of O_DIRECT by default as it results in crashes Alex Bligh
2013-04-05 10:34                                           ` [Qemu-devel] [PATCH] [RFC] Xen PV backend: Move call to bdrv_new from blk_init to blk_connect Alex Bligh
2013-04-01 16:35                                   ` [Qemu-devel] [Xen-devel] [PATCHv3] QEMU(upstream): Disable xen's use of O_DIRECT by default as it results in crashes Alex Bligh

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=45EC1ED73B198669B314A507@nimrod.local \
    --to=alex@alex.org.uk \
    --cc=Ian.Campbell@citrix.com \
    --cc=Ian.Jackson@eu.citrix.com \
    --cc=Stefano.Stabellini@eu.citrix.com \
    --cc=anthony@codemonkey.ws \
    --cc=george.dunlap@eu.citrix.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=xen-devel@lists.xen.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 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).