Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
Date:   Fri, 30 Jul 2021 08:09:32 -0700
From:   Davidlohr Bueso <dbueso@suse.de>
To:     cgel.zte@gmail.com
Cc:     keescook@chromium.org, christian.brauner@ubuntu.com,
        ktkhai@virtuozzo.com, jamorris@linux.microsoft.com,
        varad.gautam@suse.com, legion@kernel.org,
        linux-kernel@vger.kernel.org, Ran Xiaokai <ran.xiaokai@zte.com.cn>
Subject: Re: [PATCH] ipc: add set_ownership() and permissions() callbacks for
 posix mqueue sysctl
In-Reply-To: <20210729030651.536326-1-ran.xiaokai@zte.com.cn>
References: <20210729030651.536326-1-ran.xiaokai@zte.com.cn>
User-Agent: Roundcube Webmail
Message-ID: <4b67a8df7823cc14ffb69657c82562b2@suse.de>
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2021-07-28 20:06, cgel.zte@gmail.com wrote:
> This patch adds a ctl_table_set per ipc namespace, and also the
> set_ownership() and permissions() callbacks for the new ctl_table_root
> for ipc mqueue syscgtls.
                   ^^ sysctls

This makes sense to me, just some nits below.

Acked-by: Davidlohr Bueso <dbueso@suse.de>

> 
> Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> ---
...
> +static int set_permissions(struct ctl_table_header *head,
> +				struct ctl_table *table)
> +{
> +	struct ipc_namespace *ipc_ns =
> +		container_of(head->set, struct ipc_namespace, mq_set);
> +	struct user_namespace *user_ns = ipc_ns->user_ns;
> +	int mode;
> +
> +	/* Allow users with CAP_SYS_RESOURCE unrestrained access */
> +	if (ns_capable(user_ns, CAP_SYS_RESOURCE))
> +		mode = (table->mode & S_IRWXU) >> 6;
> +	else
> +	/* Allow all others at most read-only access */
> +		mode = table->mode & S_IROTH;

Please use curly braces for the else.

> +	return (mode << 6) | (mode << 3) | mode;
> +}
> +
> +static void set_ownership(struct ctl_table_header *head,
> +				struct ctl_table *table,
> +				kuid_t *uid, kgid_t *gid)
> +{
> +	struct ipc_namespace *ipc_ns =
> +		container_of(head->set, struct ipc_namespace, mq_set);
> +	struct user_namespace *user_ns = ipc_ns->user_ns;
> +	kuid_t ns_root_uid;
> +	kgid_t ns_root_gid;
> +
> +	ns_root_uid = make_kuid(user_ns, 0);
> +	if (uid_valid(ns_root_uid))
> +		*uid = ns_root_uid;
> +
> +	ns_root_gid = make_kgid(user_ns, 0);
> +	if (gid_valid(ns_root_gid))
> +		*gid = ns_root_gid;
> +}

Could set_permissions() and set_ownership() be factored such that we can 
avoid duplicated code between ipc and net ns? Something like:

void set_permissions(struct ctl_table_header *head, struct ctl_table 
*table)
{
     struct ipc_namespace *ipc_ns = container_of(head->set, struct 
ipc_namespace, mq_set);
     set_permissions_common(ipc_ns->user_ns);
}

Thanks,
Davidlohr