mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Paul Menage" <menage@google.com>
To: "Serge E. Hallyn" <serue@us.ibm.com>
Cc: "Dan Smith" <danms@us.ibm.com>,
	"Linux Containers" <containers@lists.osdl.org>,
	lkml <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH RFC] cgroup_clone: use pid of newly created task for new cgroup
Date: Wed, 11 Jun 2008 00:24:31 -0700	[thread overview]
Message-ID: <6599ad830806110024i495c5e65u82828b7237434052@mail.gmail.com> (raw)
In-Reply-To: <20080610212302.GA1946@us.ibm.com>

On Tue, Jun 10, 2008 at 2:23 PM, Serge E. Hallyn <serue@us.ibm.com> wrote:
> From faa707a44b971f5f3bf24e6a0c760ccb4ad278e6 Mon Sep 17 00:00:00 2001
> From: Serge Hallyn <serge@us.ibm.com>
> Date: Tue, 10 Jun 2008 15:57:32 -0500
> Subject: [PATCH 1/1] cgroup_clone: use pid of newly created task for new cgroup
>
> cgroup_clone creates a new cgroup with the pid of the task.  This works
> correctly for unshare, but for clone cgroup_clone is called from
> copy_namespaces inside copy_process, which happens before the new pid
> is created.  As a result, the new cgroup was created with current's pid.
> This patch:
>
>        1. Moves the call inside copy_process to after the new pid
>           is created
>        2. Passes the struct pid into ns_cgroup_clone (as it is not
>           yet attached to the task)
>        3. Passes a name from ns_cgroup_clone() into cgroup_clone()
>           so as to keep cgroup_clone() itself simpler
>        4. Uses pid_vnr() to get the process id value, so that the
>           pid used to name the new cgroup is always the pid as it
>           would be known to the task which did the cloning or
>           unsharing.  I think that is the most intuitive thing to
>           do.  This way, task t1 does clone(CLONE_NEWPID) to get
>           t2, which does clone(CLONE_NEWPID) to get t3, then the
>           cgroup for t3 will be named for the pid by which t2 knows
>           t3.
>
> This hasn't been tested enough to request inclusion, but I'd like to
> get feedback especially from Paul Menage on whether the semantics
> make sense.

Seems like a reasonable idea. It represents yet another change to the
userspace API following the 2.6.25.x one, but I guess that again it's
not one that anyone is seriously relying on yet (in particular since
it's not usable more than once from the same parent currently).

> -int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys)
> +int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys,
> +                                                               char *name)

You could reduce the patch churn by naming this parameter nodename.
> -       return cgroup_clone(task, &ns_subsys);
> +       struct pid *pid = (inpid ? inpid : task_pid(task));
> +       char name[MAX_CGROUP_TYPE_NAMELEN];

We should probably stop using MAX_CGROUP_TYPE_NAMELEN for this buffer
length and use something that explicitly sized to fit a pid_t.


> +
> +       snprintf(name, MAX_CGROUP_TYPE_NAMELEN, "%d", pid_vnr(pid));
> +       return cgroup_clone(task, &ns_subsys, name);
>  }
>
>  /*
> diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c
> index adc7851..5ca106d 100644
> --- a/kernel/nsproxy.c
> +++ b/kernel/nsproxy.c
> @@ -157,12 +157,6 @@ int copy_namespaces(unsigned long flags, struct task_struct *tsk)
>                goto out;
>        }
>
> -       err = ns_cgroup_clone(tsk);
> -       if (err) {
> -               put_nsproxy(new_ns);
> -               goto out;
> -       }
> -
>        tsk->nsproxy = new_ns;
>
>  out:
> @@ -209,7 +203,7 @@ int unshare_nsproxy_namespaces(unsigned long unshare_flags,
>                goto out;
>        }
>
> -       err = ns_cgroup_clone(current);
> +       err = ns_cgroup_clone(current, NULL);

Maybe pass task_pid(current) here rather than doing the ?: in
ns_cgroup_clone() ?

Paul

  reply	other threads:[~2008-06-11  7:24 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-06-10 21:23 Serge E. Hallyn
2008-06-11  7:24 ` Paul Menage [this message]
2008-06-11 15:46   ` Serge E. Hallyn
2008-06-11 15:59     ` Paul Menage
2008-06-12  0:37       ` Serge E. Hallyn
2008-06-11 14:30 ` Dan Smith
2008-06-11 15:46   ` Serge E. Hallyn
2008-06-11 15:49     ` Dan Smith

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6599ad830806110024i495c5e65u82828b7237434052@mail.gmail.com \
    --to=menage@google.com \
    --cc=containers@lists.osdl.org \
    --cc=danms@us.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=serue@us.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®