linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bernd Schubert <bschubert@ddn.com>
To: Miklos Szeredi <miklos@szeredi.hu>, Vivek Goyal <vgoyal@redhat.com>
Cc: Dharmendra Hans <dharamhans87@gmail.com>,
	linux-fsdevel@vger.kernel.org,
	fuse-devel <fuse-devel@lists.sourceforge.net>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 0/3] FUSE: Implement atomic lookup + open/create
Date: Wed, 11 May 2022 11:59:45 +0200	[thread overview]
Message-ID: <8ee12e2e-94b0-a66c-104c-9b8cec92b5a5@ddn.com> (raw)
In-Reply-To: <CAJfpegseGaWHkjdQj7XiR=TQNFpPZzDF_rTXces2oRz=x0N7OA@mail.gmail.com>



On 5/11/22 11:40, Miklos Szeredi wrote:
> On Thu, 5 May 2022 at 21:59, Vivek Goyal <vgoyal@redhat.com> wrote:
> 
>> Oh, I have no issues with the intent. I will like to see cut in network
>> traffic too (if we can do this without introducing problems). My primary
>> interest is that this kind of change should benefit virtiofs as well.
> 
> One issue with that appears to be checking permissions.   AFAIU this
> patchset only enables the optimization if default_permissions is
> turned off (i.e. all permission checking is done by the server).  But
> virtiofs uses the default_permissions model.
> 
> I'm not quite sure about this limitation, guessing that it's related
> to the fact that the permissions may be stale at the time of checking?

Yes exactly, I had actually pointed this out in one of the mails.

<quote myself from 2022-04-05 >
Adding in a vfs ->open_revalidate might have the advantage that we could 
also support 'default_permissions' - ->open_revalidate  needs to 
additionally check the retrieved file permissions and and needs to call 
into generic_permissions for that. Right now that is not easily 
feasible, without adding some code dup to convert flags in MAY_* flags - 
a vfs change would be needed here to pass the right flags.
</quote>


Avoiding lookup for create should work without default_permissions, though.


With the current patches this comment should be added in 
fuse_dentry_revalidate() to clarify things (somehow it got lost in the 
(internal) review process).

/* Default permissions actually also would not need revalidate, if 
atomic_open() could verify attributes after opening the file. But the 
MAY_ flags are missing and the vfs build_open_flags() to recreate these 
flags not exported. Thus, default_permissions() cannot be called from 
here - vfs updates would be required */


Thanks,
Bernd

  reply	other threads:[~2022-05-11 10:01 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-02 10:25 [PATCH v4 0/3] FUSE: Implement atomic lookup + open/create Dharmendra Singh
2022-05-02 10:25 ` [PATCH v4 1/3] FUSE: Implement atomic lookup + create Dharmendra Singh
2022-05-03 12:43   ` Vivek Goyal
2022-05-03 14:13   ` Vivek Goyal
2022-05-03 19:53   ` Vivek Goyal
2022-05-03 20:48     ` Bernd Schubert
2022-05-04  4:26     ` Dharmendra Hans
2022-05-04 14:47       ` Vivek Goyal
2022-05-04 15:46         ` Bernd Schubert
2022-05-04 17:31           ` Vivek Goyal
2022-05-05  4:51         ` Dharmendra Hans
2022-05-05 14:26           ` Vivek Goyal
2022-05-06  5:34             ` Dharmendra Hans
2022-05-06 14:12               ` Vivek Goyal
2022-05-06 16:41                 ` Bernd Schubert
2022-05-06 17:07                   ` Vivek Goyal
2022-05-06 18:45                     ` Bernd Schubert
2022-05-07 10:42                       ` Jean-Pierre André
2022-05-11 10:08                         ` Bernd Schubert
2022-05-02 10:25 ` [PATCH v4 2/3] FUSE: Implement atomic lookup + open Dharmendra Singh
2022-05-04 18:20   ` Vivek Goyal
2022-05-05  6:39     ` Dharmendra Hans
2022-05-02 10:25 ` [PATCH v4 3/3] FUSE: Avoid lookup in d_revalidate() Dharmendra Singh
2022-05-04 20:39   ` Vivek Goyal
2022-05-04 21:05     ` Bernd Schubert
2022-05-05  5:49     ` Dharmendra Hans
2022-05-04 19:18 ` [PATCH v4 0/3] FUSE: Implement atomic lookup + open/create Vivek Goyal
2022-05-05  6:12   ` Dharmendra Hans
2022-05-05 12:54     ` Vivek Goyal
2022-05-05 15:13       ` Bernd Schubert
2022-05-05 19:59         ` Vivek Goyal
2022-05-11  9:40           ` Miklos Szeredi
2022-05-11  9:59             ` Bernd Schubert [this message]
2022-05-11 17:21             ` Vivek Goyal
2022-05-11 19:30               ` Vivek Goyal
2022-05-12  8:16                 ` Dharmendra Hans
2022-05-12 15:24                   ` Vivek Goyal

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=8ee12e2e-94b0-a66c-104c-9b8cec92b5a5@ddn.com \
    --to=bschubert@ddn.com \
    --cc=dharamhans87@gmail.com \
    --cc=fuse-devel@lists.sourceforge.net \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=miklos@szeredi.hu \
    --cc=vgoyal@redhat.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).