Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Snaipe <snaipe@arista.com>
To:     hughd@google.com
Cc:     ovt@google.com, corbet@lwn.net, akpm@linux-foundation.org,
        brauner@kernel.org, linux-doc@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        Snaipe <snaipe@arista.com>
Subject: Re: [PATCH] shmem: add support for user extended attributes
Date:   Mon, 14 Aug 2023 10:23:39 +0200
Message-ID: <20230814082339.2006418-1-snaipe@arista.com>
In-Reply-To: <9b8d38f0-fd22-3f98-d070-16baf976ecb5@google.com>
References: <9b8d38f0-fd22-3f98-d070-16baf976ecb5@google.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

> Your sending this patch does help to raise the priority for my
> sending that patch: thank you; but I cannot promise when that will be.

Hey Hugh,

Just as an additional data point, if it helps with priority :)

The lack of support of user xattrs on tmpfs our last remaining blocker for
using unprivileged overlayfs mounts that use a tmpfs for the upper dir & work
dir. Not that it isn't possible to use unprivileged overlayfs mounts in general,
but not having this option means that use-cases for discardable upper layer
changes are harder to clean up correctly when not on an tmpfs mount whose
lifetime is bound to a mount namespace.

I don't think there's any rush; we can live with rmdir failing with EIO for now,
but it'd be great to see this fixed rather than having to implement expensive
cleanup routines that have to remove the upper+work dirs recursively with
CAP_DAC_OVERRIDE.

Cheers,

-- 
Snaipe