Re: [PATCH v2] ovl: require xwhiteout feature flag on layer roots

From: Miklos Szeredi <miklos@szeredi.hu>
To: Amir Goldstein <amir73il@gmail.com>
Cc: Alexander Larsson <alexl@redhat.com>,
	Miklos Szeredi <mszeredi@redhat.com>,
	linux-unionfs@vger.kernel.org,  linux-fsdevel@vger.kernel.org,
	stable@vger.kernel.org
Subject: Re: [PATCH v2] ovl: require xwhiteout feature flag on layer roots
Date: Fri, 19 Jan 2024 21:29:59 +0100	[thread overview]
Message-ID: <CAJfpeguFY8KX9kXPBgz5imVTV4A0R+aqS_SRiwdoPXPqR_B_xg@mail.gmail.com> (raw)
In-Reply-To: <CAOQ4uxiob0t4YDpEZ4urfro=NrXF+FH_Bvt9DbD1cHbJAWf88A@mail.gmail.com>

On Fri, 19 Jan 2024 at 20:06, Amir Goldstein <amir73il@gmail.com> wrote:

> How about checking xwhiteouts xattrs along with impure and
> origin xattrs in ovl_get_inode()?
>
> Then there will be no overhead in readdir and no need for
> marking the layer root?
>
> Miklos, would that be acceptable?

It's certainly a good idea, but doesn't really address my worry.  The
minor performance impact is not what bothers me most.  It's the fact
that in the common case the result of these calls are discarded.
That's just plain ugly, IMO.

My preferred alternative would be a mount option.  Amir, Alex, would
you both be okay with that?

Thanks,
Miklos