Received-SPF: pass (google.com: domain of linux-ext4+bounces-510-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Thu, 21 Dec 2023 07:14:02 +0000
From: Al Viro <viro@zeniv.linux.org.uk>
To: Gabriel Krisman Bertazi <krisman@suse.de>
Cc: ebiggers@kernel.org, jaegeuk@kernel.org, tytso@mit.edu,
	linux-f2fs-devel@lists.sourceforge.net, linux-ext4@vger.kernel.org,
	linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH v2 2/8] fscrypt: Drop d_revalidate if key is available
Message-ID: <20231221071402.GA1674809@ZenIV>
References: <20231215211608.6449-1-krisman@suse.de>
 <20231215211608.6449-3-krisman@suse.de>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20231215211608.6449-3-krisman@suse.de>
Sender: Al Viro <viro@ftp.linux.org.uk>

On Fri, Dec 15, 2023 at 04:16:02PM -0500, Gabriel Krisman Bertazi wrote:
> fscrypt dentries are always valid once the key is available.  Since the
> key cannot be removed without evicting the dentry, we don't need to keep
> retrying to revalidate it.
> 
> Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
> ---
>  fs/crypto/fname.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/crypto/fname.c b/fs/crypto/fname.c
> index 7b3fc189593a..0457ba2d7d76 100644
> --- a/fs/crypto/fname.c
> +++ b/fs/crypto/fname.c
> @@ -591,8 +591,15 @@ int fscrypt_d_revalidate(struct dentry *dentry, unsigned int flags)
>  	 * reverting to no-key names without evicting the directory's inode
>  	 * -- which implies eviction of the dentries in the directory.
>  	 */
> -	if (!(dentry->d_flags & DCACHE_NOKEY_NAME))
> +	if (!(dentry->d_flags & DCACHE_NOKEY_NAME)) {
> +		/*
> +		 * If fscrypt is the only feature requiring
> +		 * revalidation for this dentry, we can just disable it.
> +		 */
> +		if (dentry->d_op->d_revalidate == &fscrypt_d_revalidate)

Umm...  What about ceph?  IOW, why do we care how had we gotten to that
function - directly via ->d_revalidate() or from ->d_revalidate() instance?