cgroup_clone: use pid of newly created task for new cgroup

From faa707a44b971f5f3bf24e6a0c760ccb4ad278e6 Mon Sep 17 00:00:00 2001
From: Serge Hallyn <[email protected]>
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.

kernel/ns_cgroup.c +3 -0

+
+       snprintf(name, MAX_CGROUP_TYPE_NAMELEN, "%d", pid_vnr(pid));
+       return cgroup_clone(task, &ns_subsys, name);
 }
 
 /*

kernel/nsproxy.c +0 -6

index adc7851..5ca106d 100644
@@ -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:

Quoted 26 lines · Serge E. Hallyn

From faa707a44b971f5f3bf24e6a0c760ccb4ad278e6 Mon Sep 17 00:00:00 2001
From: Serge Hallyn <[email protected]>
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.

Quoted 3 lines

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).

Review

Quoted 3 lines

(quoted diff) +2 -1

-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.

Review

Quoted 3 lines

(quoted diff) +2 -1

-       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.

Quoted 27 lines

(quoted diff) +3 -0

+
+       snprintf(name, MAX_CGROUP_TYPE_NAMELEN, "%d", pid_vnr(pid));
+       return cgroup_clone(task, &ns_subsys, name);
 }
 
 /*

kernel/nsproxy.c +0 -6

index adc7851..5ca106d 100644
@@ -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;
       }
 

kernel/nsproxy.c int unshare_nsproxy_namespaces(unsigned long unshare_flags, · line 206

Quoted 2 lines

kernel/nsproxy.c +1 -1

-       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

From f0635c20e9e9643fa9a90dd7e29b7855ff32ad40 Mon Sep 17 00:00:00 2001
From: Serge Hallyn <[email protected]>
Date: Wed, 11 Jun 2008 10:41:37 -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.

(Thanks to Dan Smith for finding the main bug)

Changelog:
June 11: Incorporate Paul Menage's feedback: don't pass
NULL to ns_cgroup_clone from unshare, and reduce patch size by using 'nodename' in cgroup_clone.
June 10: Original version

Signed-off-by Serge Hallyn <[email protected]>

---
 include/linux/cgroup.h  |    3 ++-
 include/linux/nsproxy.h |    7 +++++--
 kernel/cgroup.c         |    7 +++----
 kernel/fork.c           |    4 ++++
 kernel/ns_cgroup.c      |    7 +++++--
 kernel/nsproxy.c        |    8 +-------
 6 files changed, 20 insertions(+), 16 deletions(-)

include/linux/cgroup.h +2 -1

index 3690c3b..72cb6ca 100644
@@ -349,7 +349,8 @@ static inline struct cgroup* task_cgroup(struct task_struct *task,
       return task_subsys_state(task, subsys_id)->cgroup;
 }
 
-int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *ss);
+int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *ss,
+                                                       char *nodename);
 
 /* A cgroup_iter should be treated as an opaque object */
 struct cgroup_iter {

include/linux/nsproxy.h +5 -2

index 0e66b57..c8a768e 100644
@@ -82,9 +82,12 @@ static inline void get_nsproxy(struct nsproxy *ns)
 }
 
 #ifdef CONFIG_CGROUP_NS
-int ns_cgroup_clone(struct task_struct *tsk);
+int ns_cgroup_clone(struct task_struct *tsk, struct pid *pid);
 #else
-static inline int ns_cgroup_clone(struct task_struct *tsk) { return 0; }
+static inline int ns_cgroup_clone(struct task_struct *tsk, struct pid *pid)
+{
+       return 0;
+}
 #endif
 
 #endif

kernel/cgroup.c +3 -4

index 79fa060..b4c4b75 100644
@@ -2869,16 +2869,17 @@ void cgroup_exit(struct task_struct *tsk, int run_callbacks)
 * cgroup_clone - clone the cgroup the given subsystem is attached to
 * @tsk: the task to be moved
 * @subsys: the given subsystem
+ * @nodename: the name for the new cgroup
 *
 * Duplicate the current cgroup in the hierarchy that the given
 * subsystem is attached to, and move this task into the new
 * child.
 */
-int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys)
+int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys,
+                                                       char *nodename)
 {
       struct dentry *dentry;
       int ret = 0;
-       char nodename[MAX_CGROUP_TYPE_NAMELEN];
       struct cgroup *parent, *child;
       struct inode *inode;
       struct css_set *cg;

@@ -2903,8 +2904,6 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys)
       cg = tsk->cgroups;
       parent = task_cgroup(tsk, subsys->subsys_id);
 
-       snprintf(nodename, MAX_CGROUP_TYPE_NAMELEN, "%d", tsk->pid);
-
       /* Pin the hierarchy */
       atomic_inc(&parent->root->sb->s_active);
 

kernel/fork.c +4 -0

index f0e7767..47c0a97 100644
@@ -1125,6 +1125,10 @@ static struct task_struct *copy_process(unsigned long clone_flags,
       if (clone_flags & CLONE_THREAD)
               p->tgid = current->tgid;
 
+       if (current->nsproxy != p->nsproxy)
+               if ((retval = ns_cgroup_clone(p, pid)))
+                       goto bad_fork_free_pid;
+
       p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL;
       /*
        * Clear TID on mm_release()?

kernel/ns_cgroup.c +5 -2

index 48d7ed6..6431fb7 100644
@@ -24,9 +24,12 @@ static inline struct ns_cgroup *cgroup_to_ns(
                           struct ns_cgroup, css);
 }
 
-int ns_cgroup_clone(struct task_struct *task)
+int ns_cgroup_clone(struct task_struct *task, struct pid *pid)
 {
-       return cgroup_clone(task, &ns_subsys);
+       char name[PROC_NUMBUF];
+
+       snprintf(name, PROC_NUMBUF, "%d", pid_vnr(pid));
+       return cgroup_clone(task, &ns_subsys, name);
 }
 
 /*

kernel/nsproxy.c +1 -8

index adc7851..21575fc 100644
@@ -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, task_pid(current));
       if (err)
               put_nsproxy(*new_nsp);
 
--

1.5.4.3

Quoted 27 lines · Serge E. Hallyn

From f0635c20e9e9643fa9a90dd7e29b7855ff32ad40 Mon Sep 17 00:00:00 2001
From: Serge Hallyn <[email protected]>
Date: Wed, 11 Jun 2008 10:41:37 -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.

Quoted 1 line

(Thanks to Dan Smith for finding the main bug)

SH> As a result, the new cgroup was created with current's pid.

This works for me and results in the creation of a cgroup named after the new task's pid instead of the parent's pid.

Thanks!