Re: 2.6.23-rc6-mm1 - make access to tasks nsproxy ligther (fix)

[Cedric Le Goater <clg@fr.ibm.com>]

Hello !

Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.23-rc6/2.6.23-rc6-mm1/

make-access-to-tasks-nsproxy-lighter.patch breaks unshare()

when called from unshare(), switch_task_namespaces() takes an 
extra refcount on the nsproxy, leading to a memory leak of 
nsproxy objects. 

Now the problem is that we still need that extra ref when called 
from daemonize(). Here's an ugly fix for it.

Signed-off-by: Cedric Le Goater <clg@fr.ibm.com>
---
 include/linux/nsproxy.h |    5 +++++
 kernel/exit.c           |    2 ++
 kernel/nsproxy.c        |    7 -------
 3 files changed, 7 insertions(+), 7 deletions(-)

Index: 2.6.23-rc6-mm1/kernel/nsproxy.c
===================================================================
--- 2.6.23-rc6-mm1.orig/kernel/nsproxy.c
+++ 2.6.23-rc6-mm1/kernel/nsproxy.c
@@ -25,11 +25,6 @@ static struct kmem_cache *nsproxy_cachep
 
 struct nsproxy init_nsproxy = INIT_NSPROXY(init_nsproxy);
 
-static inline void get_nsproxy(struct nsproxy *ns)
-{
-	atomic_inc(&ns->count);
-}
-
 /*
  * creates a copy of "orig" with refcount 1.
  */
@@ -208,8 +203,6 @@ void switch_task_namespaces(struct task_
 	if (ns == new)
 		return;
 
-	if (new)
-		get_nsproxy(new);
 	rcu_assign_pointer(p->nsproxy, new);
 
 	if (ns && atomic_dec_and_test(&ns->count)) {
Index: 2.6.23-rc6-mm1/kernel/exit.c
===================================================================
--- 2.6.23-rc6-mm1.orig/kernel/exit.c
+++ 2.6.23-rc6-mm1/kernel/exit.c
@@ -408,6 +408,8 @@ void daemonize(const char *name, ...)
 	current->fs = fs;
 	atomic_inc(&fs->count);
 
+	if (current->nsproxy != init_task.nsproxy)
+		get_nsproxy(init_task.nsproxy);
 	switch_task_namespaces(current, init_task.nsproxy);
 
 	exit_files(current);
Index: 2.6.23-rc6-mm1/include/linux/nsproxy.h
===================================================================
--- 2.6.23-rc6-mm1.orig/include/linux/nsproxy.h
+++ 2.6.23-rc6-mm1/include/linux/nsproxy.h
@@ -77,6 +77,11 @@ static inline void put_nsproxy(struct ns
 	}
 }
 
+static inline void get_nsproxy(struct nsproxy *ns)
+{
+	atomic_inc(&ns->count);
+}
+
 #ifdef CONFIG_CONTAINER_NS
 int ns_container_clone(struct task_struct *tsk);
 #else
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/