linux-unionfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Vivek Goyal <vgoyal@redhat.com>
To: Amir Goldstein <amir73il@gmail.com>
Cc: Miklos Szeredi <miklos@szeredi.hu>, linux-unionfs@vger.kernel.org
Subject: Re: [PATCH] ovl: fix lookup of indexed hardlinks with metacopy
Date: Thu, 16 Jul 2020 09:22:20 -0400	[thread overview]
Message-ID: <20200716132220.GA422759@redhat.com> (raw)
In-Reply-To: <20200715133808.7146-1-amir73il@gmail.com>

On Wed, Jul 15, 2020 at 04:38:08PM +0300, Amir Goldstein wrote:
> We recently moved setting inode flag OVL_UPPERDATA to ovl_lookup().
> 
> When looking up an overlay dentry, upperdentry may be found by index
> and not by name.  In that case, we fail to read the metacopy xattr
> and falsly set the OVL_UPPERDATA on the overlay inode.
> 
> This caused a regression in xfstest overlay/033 when run with
> OVERLAY_MOUNT_OPTIONS="-o metacopy=on".
> 
> Fixes: 28166ab3c875 ("ovl: initialize OVL_UPPERDATA in ovl_lookup()")
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>

Reviewed-by: Vivek Goyal <vgoyal@redhat.com>

Good catch Amir. Thanks.


> ---
> 
> Miklos,
> 
> I just ran xfstests -g overlay/quick tests with metacopy enabled
> and one test failed.
> 
> Vivek,
> 
> Do you by any chance run this sort of test regularly?

No I don't. I will start testing this configuration. Actually I don't
have any automated setup and all the testing I do manually. I think
its time that atleast I have a script which runs bunch of tests
always.

> 
> You have asked about running unionmount tests with metacopy before.
> I just pushed a commit to my xfstests 'unionmount' branch:
>   7859f22b ovl: test unionmount tests with metacopy
> 
> It allows you to run xfstests -g overlay/union with
> OVERLAY_MOUNT_OPTIONS="-o metacopy=on", to excercise all the unionmount
> test configurations I created with metacopy enabled.

Nice, I will check it out. Will be nice if this gets pushed to
xfstest tree.

Thanks
Vivek

> 
> Maybe not so surprising that the sub-group overlay/union.nested tests
> fail with metacopy enabled. This is just a test setup bug and I pushed
> a fix commit to unionmount overlayfs-devel branch to fix it.
> 
> Thanks,
> Amir.
> 
>  fs/overlayfs/namei.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/fs/overlayfs/namei.c b/fs/overlayfs/namei.c
> index 3566282a9199..0c5a624600c1 100644
> --- a/fs/overlayfs/namei.c
> +++ b/fs/overlayfs/namei.c
> @@ -1073,6 +1073,10 @@ struct dentry *ovl_lookup(struct inode *dir, struct dentry *dentry,
>  			upperredirect = NULL;
>  			goto out_free_oe;
>  		}
> +		err = ovl_check_metacopy_xattr(upperdentry);
> +		if (err < 0)
> +			goto out_free_oe;
> +		uppermetacopy = err;
>  	}
>  
>  	if (upperdentry || ctr) {
> -- 
> 2.17.1
> 


      reply	other threads:[~2020-07-16 13:22 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-15 13:38 [PATCH] ovl: fix lookup of indexed hardlinks with metacopy Amir Goldstein
2020-07-16 13:22 ` Vivek Goyal [this message]

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=20200716132220.GA422759@redhat.com \
    --to=vgoyal@redhat.com \
    --cc=amir73il@gmail.com \
    --cc=linux-unionfs@vger.kernel.org \
    --cc=miklos@szeredi.hu \
    /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).