Re: [RFC PATCH 5 7/7] KEYS: exec request key within service thread of key creator

From: David Howells <dhowells@redhat.com>
To: Ian Kent <raven@themaw.net>
Cc: dhowells@redhat.com,
	Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Oleg Nesterov <onestero@redhat.com>,
	Trond Myklebust <trond.myklebust@primarydata.com>,
	"J. Bruce Fields" <bfields@fieldses.org>,
	Benjamin Coddington <bcodding@redhat.com>,
	Al Viro <viro@ZenIV.linux.org.uk>,
	Jeff Layton <jeff.layton@primarydata.com>,
	"Eric W. Biederman" <ebiederm@xmission.com>
Subject: Re: [RFC PATCH 5 7/7] KEYS: exec request key within service thread of key creator
Date: Thu, 02 Apr 2015 13:58:04 +0100	[thread overview]
Message-ID: <18931.1427979484@warthog.procyon.org.uk> (raw)
In-Reply-To: <20150331031533.10464.1615.stgit@pluto.fritz.box>

Ian Kent <raven@themaw.net> wrote:

> +
> +	/* Namespace token */
> +	int umh_token;

If you could put it after data_len so that all the smaller-than-wordsize
fields are together for better packing.

> +		umh_wq_put_token(key->umh_token);

Does gc.c need an extra #include for this?

> +	/* If running within a container use the container namespace */
> +	if (current->nsproxy->net_ns != &init_net)
> +		key->umh_token = umh_wq_get_token(0, "keys");

So keys live in the networking namespace?

> -	ret = call_usermodehelper_keys(argv[0], argv, envp, keyring,
> -				       UMH_WAIT_PROC);
> +	/* If running within a container use the container namespace */
> +	if (key->umh_token)
> +		ret = call_usermodehelper_keys_service(argv[0], argv, envp,
> +						       keyring, key->umh_token,
> +						       UMH_WAIT_PROC);
> +	else
> +		ret = call_usermodehelper_keys(argv[0], argv, envp,
> +					       keyring, UMH_WAIT_PROC);

call_usermodehelper_keys_service() would appear to be superfluous.  If
key->umh_token is 0, you call call_usermodehelper_keys() which then calls
call_usermodehelper_keys_service() with a 0 token...

David