qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eric Blake <eblake@redhat.com>
To: Supriya Kannery <supriyak@linux.vnet.ibm.com>
Cc: Kevin Wolf <kwolf@redhat.com>,
	Shrinidhi Joshi <spjoshi31@gmail.com>,
	Stefan Hajnoczi <stefanha@gmail.com>,
	Jeff Cody <jcody@redhat.com>,
	qemu-devel@nongnu.org, Luiz Capitulino <lcapitulino@redhat.com>,
	Christoph Hellwig <hch@lst.de>
Subject: Re: [Qemu-devel] [v1 Patch 4/10]Qemu: Framework for reopening image files safely
Date: Fri, 15 Jun 2012 16:02:30 -0600	[thread overview]
Message-ID: <4FDBB0F6.6000407@redhat.com> (raw)
In-Reply-To: <20120615204744.9853.62830.sendpatchset@skannery.in.ibm.com>

[-- Attachment #1: Type: text/plain, Size: 2361 bytes --]

On 06/15/2012 02:47 PM, Supriya Kannery wrote:
> Struct BDRVReopenState along with three reopen related functions
> introduced for handling reopening of images safely. This can be
> extended by each of the block drivers to reopen respective
> image files.
> 
> Signed-off-by: Supriya Kannery <supriyak@linux.vnet.ibm.com>
> 

> +    /* Use driver specific reopen() if available */
> +    if (drv->bdrv_reopen_prepare) {
> +        ret = bdrv_reopen_prepare(bs, &reopen_state, bdrv_flags);
> +         if (ret < 0) {
> +            bdrv_reopen_abort(bs, reopen_state);
> +            qerror_report(QERR_OPEN_FILE_FAILED, bs->filename);
> +            return ret;
> +        }
> +
> +        bdrv_reopen_commit(bs, reopen_state);
> +        bs->open_flags = bdrv_flags;

Good.

> +
> +    } else {
> +
> +        /* Try reopening the image using protocol or directly */
> +        if (bs->file) {
> +            open_flags = bs->open_flags;
> +            drv->bdrv_close(bs);
> +            if (drv->bdrv_file_open) {
> +                ret = drv->bdrv_file_open(bs, bs->filename, bdrv_flags);

Not so good.  Are we sure we can't flush all data, so that we can then
attempt the open before the close, and avoid losing the original file on
error?

> +            } else {
> +                ret = bdrv_file_open(&bs->file, bs->filename, bdrv_flags);
> +            }
> +            if (ret < 0) {
> +                /* Reopen failed. Try to open with original flags */
> +                qerror_report(QERR_OPEN_FILE_FAILED, bs->filename);
> +                if (drv->bdrv_file_open) {
> +                    ret = drv->bdrv_file_open(bs, bs->filename, open_flags);
> +                } else {
> +                    ret = bdrv_file_open(&bs->file, bs->filename, open_flags);
> +                }
> +                if (ret < 0) {
> +                    /* Reopen failed with orig and modified flags */
> +                    bdrv_delete(bs->file);
> +                    bs->drv = NULL;

You dropped the abort() here, but now the device is silently gone.  But
since the error of QERR_OPEN_FILE_FAILED is the same as in the safe
case, how does the management app know that things were corrupted?

-- 
Eric Blake   eblake@redhat.com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 620 bytes --]

  reply	other threads:[~2012-06-15 22:02 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-06-15 20:46 [Qemu-devel] [v1 Patch 0/10]Qemu: Dynamic host pagecache change and image file reopen Supriya Kannery
2012-06-15 20:47 ` [Qemu-devel] [v1 Patch 1/10]Qemu: Enhance "info block" to display host cache setting Supriya Kannery
2012-06-15 21:07   ` Eric Blake
2012-07-09 14:43     ` Kevin Wolf
2012-07-11 14:03       ` Luiz Capitulino
2012-07-29  6:21         ` Supriya Kannery
2012-07-05 16:38   ` Jeff Cody
2012-07-29  6:54     ` Supriya Kannery
2012-06-15 20:47 ` [Qemu-devel] [v1 Patch 2/10]Qemu: Error classes for hostcache setting and data sync failures Supriya Kannery
2012-07-09 14:47   ` Kevin Wolf
2012-07-29  6:58     ` Supriya Kannery
2012-06-15 20:47 ` [Qemu-devel] [v1 Patch 3/10]Qemu: Cmd "block_set_hostcache" for dynamic cache change Supriya Kannery
2012-06-15 21:56   ` Eric Blake
2012-07-29  7:33     ` Supriya Kannery
2012-06-20 18:18   ` Jeff Cody
2012-07-04  5:10     ` Shrinidhi Joshi
2012-07-04  6:30       ` Kevin Wolf
2012-07-09 14:52   ` Kevin Wolf
2012-07-11 14:16   ` Luiz Capitulino
2012-07-29  7:56     ` Supriya Kannery
2012-06-15 20:47 ` [Qemu-devel] [v1 Patch 4/10]Qemu: Framework for reopening image files safely Supriya Kannery
2012-06-15 22:02   ` Eric Blake [this message]
2012-07-09 15:06   ` Kevin Wolf
2012-06-15 20:48 ` [Qemu-devel] [v1 Patch 5/10]Qemu: raw-posix image file reopen Supriya Kannery
2012-06-15 22:11   ` Eric Blake
2012-07-04  5:15     ` Shrinidhi Joshi
2012-07-04 11:32       ` Eric Blake
2012-06-15 20:48 ` [Qemu-devel] [v1 Patch 6/10]Qemu: raw-win32 " Supriya Kannery
2012-06-15 20:48 ` [Qemu-devel] [v1 Patch 7/10]Qemu: vmdk " Supriya Kannery
2012-06-15 20:48 ` [Qemu-devel] [v1 Patch 8/10]Qemu: qcow2 " Supriya Kannery
2012-06-15 20:48 ` [Qemu-devel] [v1 Patch 9/10]Qemu: qcow " Supriya Kannery
2012-06-15 20:49 ` [Qemu-devel] [v1 Patch 10/10]Qemu: qed " Supriya Kannery
2012-07-09 17:51 ` [Qemu-devel] [v1 Patch 0/10]Qemu: Dynamic host pagecache change and " Stefan Weil

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=4FDBB0F6.6000407@redhat.com \
    --to=eblake@redhat.com \
    --cc=hch@lst.de \
    --cc=jcody@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=lcapitulino@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=spjoshi31@gmail.com \
    --cc=stefanha@gmail.com \
    --cc=supriyak@linux.vnet.ibm.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).