From: Anthony Liguori <aliguori@us.ibm.com>
To: Michael Roth <mdroth@linux.vnet.ibm.com>, qemu-devel@nongnu.org
Cc: quintela@redhat.com, owasserm@redhat.com, yamahata@valinux.co.jp,
pbonzini@redhat.com, akong@redhat.com, afaerber@suse.de
Subject: Re: [Qemu-devel] [PATCH 12/22] qom-fuse: workaround for truncated properties > 4096
Date: Tue, 24 Jul 2012 17:26:43 -0500 [thread overview]
Message-ID: <87pq7kkd7g.fsf@codemonkey.ws> (raw)
In-Reply-To: <1343150454-4677-13-git-send-email-mdroth@linux.vnet.ibm.com>
Michael Roth <mdroth@linux.vnet.ibm.com> writes:
I don't think this is the most correct solution but I think it's good
enough in practice so
Reviewed-by: Anthony Liguori <aliguori@us.ibm.com>
Regards,
Anthony Liguori
> We currently hard-code property size at 4096 for the purposes of
> getattr()/stat()/etc. For 'state' properties we can exceed this easily,
> leading to truncated responses.
>
> Instead, for a particular property, make it
> max(4096, most_recent_property_size * 2). This allows some
> head-room for properties that change size periodically (numbers,
> strings, state properties containing arrays, etc)
>
> Also, implement a simple property cache to avoid spinning on qom-get
> if an application reads beyond the actual size. This also allows us
> to use a snapshot of a single qom-get that persists across read()'s.
> Old Cache entries are evicted as soon as we attempt to read() from
> offset 0 again.
>
> Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
> ---
> QMP/qom-fuse | 24 +++++++++++++++++-------
> 1 file changed, 17 insertions(+), 7 deletions(-)
>
> diff --git a/QMP/qom-fuse b/QMP/qom-fuse
> index 5c6754a..bd43f29 100755
> --- a/QMP/qom-fuse
> +++ b/QMP/qom-fuse
> @@ -26,6 +26,7 @@ class QOMFS(Fuse):
> self.qmp.connect()
> self.ino_map = {}
> self.ino_count = 1
> + self.prop_cache = {}
>
> def get_ino(self, path):
> if self.ino_map.has_key(path):
> @@ -67,12 +68,16 @@ class QOMFS(Fuse):
> if not self.is_property(path):
> return -ENOENT
>
> - path, prop = path.rsplit('/', 1)
> - try:
> - data = str(self.qmp.command('qom-get', path=path, property=prop))
> - data += '\n' # make values shell friendly
> - except:
> - return -EPERM
> + # avoid extra calls to qom-get by using cached value when offset > 0
> + if offset == 0 or not self.prop_cache.has_key(path):
> + directory, prop = path.rsplit('/', 1)
> + try:
> + resp = str(self.qmp.command('qom-get', path=directory, property=prop))
> + self.prop_cache[path] = resp + '\n' # make values shell friendly
> + except:
> + return -EPERM
> +
> + data = self.prop_cache[path]
>
> if offset > len(data):
> return ''
> @@ -111,13 +116,18 @@ class QOMFS(Fuse):
> 0,
> 0))
> elif self.is_property(path):
> + directory, prop = path.rsplit('/', 1)
> + try:
> + resp = str(self.qmp.command('qom-get', path=directory, property=prop))
> + except:
> + return -ENOENT
> value = posix.stat_result((0644 | stat.S_IFREG,
> self.get_ino(path),
> 0,
> 1,
> 1000,
> 1000,
> - 4096,
> + max(len(resp) * 2, 4096),
> 0,
> 0,
> 0))
> --
> 1.7.9.5
next prev parent reply other threads:[~2012-07-24 22:26 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-07-24 17:20 [Qemu-devel] [RFC v2] Use QEMU IDL for device serialization/introspection Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 01/22] qapi: qapi-visit.py -> qapi_visit.py so we can import Michael Roth
2012-07-24 22:12 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 02/22] qapi: qapi-types.py -> qapi_types.py Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 03/22] qapi: qapi-commands.py -> qapi_commands.py Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 04/22] qapi: qapi_visit.py, make code useable as module Michael Roth
2012-07-24 22:13 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 05/22] qapi: qapi_visit.py, support arrays and complex qapi definitions Michael Roth
2012-07-24 22:16 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 06/22] qapi: add visitor interfaces for arrays Michael Roth
2012-07-24 22:18 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 07/22] qapi: qapi_visit.py, support generating static functions Michael Roth
2012-07-24 22:19 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 08/22] qapi: qapi_visit.py, support for visiting non-pointer/embedded structs Michael Roth
2012-07-24 22:21 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 09/22] qapi: QmpOutputVisitor, implement array handling Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 10/22] qapi: qapi.py, make json parser more robust Michael Roth
2012-07-24 22:23 ` Anthony Liguori
2012-07-24 22:59 ` Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 11/22] qapi: add open-coded visitor for struct tm types Michael Roth
2012-07-24 22:24 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 12/22] qom-fuse: workaround for truncated properties > 4096 Michael Roth
2012-07-24 22:26 ` Anthony Liguori [this message]
2012-07-24 17:20 ` [Qemu-devel] [PATCH 13/22] module additions for schema registration Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 14/22] qdev: move Property-related declarations to qdev-properties.h Michael Roth
2012-07-24 22:28 ` Anthony Liguori
2012-07-24 17:20 ` [Qemu-devel] [PATCH 15/22] qidl: Add documentation Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 16/22] qidl: parser, initial import from qc.git Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 17/22] qidl: codegen, initial commit Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 18/22] qidl: qidl.h Michael Roth
2012-07-24 20:47 ` Blue Swirl
2012-07-25 0:30 ` Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 19/22] qidl: unit tests Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 20/22] qemu-timer: add visit_type_QEMUTimer Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 21/22] rtc: add QIDL annotations Michael Roth
2012-07-24 17:20 ` [Qemu-devel] [PATCH 22/22] cirrus_vga: " Michael Roth
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=87pq7kkd7g.fsf@codemonkey.ws \
--to=aliguori@us.ibm.com \
--cc=afaerber@suse.de \
--cc=akong@redhat.com \
--cc=mdroth@linux.vnet.ibm.com \
--cc=owasserm@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
--cc=yamahata@valinux.co.jp \
/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).