linux-kernel - Re: [PATCH v2] vfs: Make vfs_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Y17edxOrW81EBh1v@ZenIV>
Date:   Sun, 30 Oct 2022 20:28:39 +0000
From:   Al Viro <viro@...iv.linux.org.uk>
To:     Dawei Li <set_pte_at@...look.com>
Cc:     brauner@...nel.org, neilb@...e.de, linux-fsdevel@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] vfs: Make vfs_get_super() internal

On Sun, Oct 30, 2022 at 09:06:08PM +0800, Dawei Li wrote:
> For now there are no external callers of vfs_get_super(),
> so just make it an internal API.
> 
> v1: https://lore.kernel.org/all/TYCP286MB2323D37F4F6400FD07D7C7F7CA319@TYCP286MB2323.JPNP286.PROD.OUTLOOK.COM/
> 
> v2: move vfs_get_super_keying to super.c, as the suggestion
> from Christian Brauner.
> 
> base-commit: 3aca47127a646165965ff52803e2b269eed91afc
> 
> Signed-off-by: Dawei Li <set_pte_at@...look.com>
> ---
>  fs/super.c                 | 13 +++++++++++--
>  include/linux/fs_context.h | 14 --------------
>  2 files changed, 11 insertions(+), 16 deletions(-)
> 
> diff --git a/fs/super.c b/fs/super.c
> index 6a82660e1adb..24e31e458552 100644
> --- a/fs/super.c
> +++ b/fs/super.c
> @@ -1111,6 +1111,16 @@ static int test_single_super(struct super_block *s, struct fs_context *fc)
>  	return 1;
>  }
>  
> +/*
> + * sget() wrappers to be called from the ->get_tree() op.
> + */
> +enum vfs_get_super_keying {
> +	vfs_get_single_super,	/* Only one such superblock may exist */
> +	vfs_get_single_reconf_super, /* As above, but reconfigure if it exists */
> +	vfs_get_keyed_super,	/* Superblocks with different s_fs_info keys may exist */
> +	vfs_get_independent_super, /* Multiple independent superblocks may exist */
> +};

I would rather kill the "keying" thing completely.

Seriously, what does it buy us?  Consider e.g.
        return vfs_get_super(fc, vfs_get_independent_super, fill_super);
in get_tree_nodev().  If you expand vfs_get_super() there, you'll get
this:
{
        struct super_block *sb = sget_fc(fc, NULL, set_anon_super_fc);
        if (IS_ERR(sb))
                return PTR_ERR(sb);

        if (!sb->s_root) {
                int err = fill_super(sb, fc);
                if (unlikely(err)) {
                        deactivate_locked_super(sb);
			return err;
		}
                sb->s_flags |= SB_ACTIVE;
        }
	fc->root = dget(sb->s_root);
	return 0;
}

Sure, it's several lines long, but it's much easier to follow than
vfs_get_super().  And duplication between these vfs_get_super()
callers is, IMO, not worth eliminating...