OpenVZ Forum


Home » Mailing lists » Devel » [PATCH 1/2] Don't forget to unlock uids_mutex on error paths
[PATCH 1/2] Don't forget to unlock uids_mutex on error paths [message #23635] Wed, 21 November 2007 10:49 Go to next message
Pavel Emelianov is currently offline  Pavel Emelianov
Messages: 1149
Registered: September 2006
Senior Member
The commit

commit 5cb350baf580017da38199625b7365b1763d7180
Author: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Date:   Mon Oct 15 17:00:14 2007 +0200

    sched: group scheduling, sysfs tunables

introduced the uids_mutex and the helpers to lock/unlock it.
Unfortunately, the error paths of alloc_uid() were not patched
to unlock it.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>

---

diff --git a/kernel/user.c b/kernel/user.c
index 56327ee..3549c4b 100644
--- a/kernel/user.c
+++ b/kernel/user.c
@@ -343,8 +343,11 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 		struct user_struct *new;
 
 		new = kmem_cache_alloc(uid_cachep, GFP_KERNEL);
-		if (!new)
+		if (!new) {
+			uids_mutex_unlock();
 			return NULL;
+		}
+
 		new->uid = uid;
 		atomic_set(&new->__count, 1);
 		atomic_set(&new->processes, 0);
@@ -361,6 +364,7 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 
 		if (alloc_uid_keyring(new, current) < 0) {
 			kmem_cache_free(uid_cachep, new);
+			uids_mutex_unlock();
 			return NULL;
 		}
 
@@ -368,6 +372,7 @@ struct user_struct * alloc_uid(struct user_namespace *ns, uid_t uid)
 			key_put(new->uid_keyring);
 			key_put(new->session_keyring);
 			kmem_cache_free(uid_cachep, new);
+			uids_mutex_unlock();
 			return NULL;
 		}
 
-- 
1.5.3.4
Re: [PATCH 1/2] Don't forget to unlock uids_mutex on error paths [message #23728 is a reply to message #23635] Fri, 23 November 2007 13:12 Go to previous messageGo to next message
Dhaval Giani is currently offline  Dhaval Giani
Messages: 37
Registered: June 2007
Member
On Wed, Nov 21, 2007 at 01:49:33PM +0300, Pavel Emelyanov wrote:
> The commit
> 
> commit 5cb350baf580017da38199625b7365b1763d7180
> Author: Dhaval Giani <dhaval@linux.vnet.ibm.com>
> Date:   Mon Oct 15 17:00:14 2007 +0200
> 
>     sched: group scheduling, sysfs tunables
> 
> introduced the uids_mutex and the helpers to lock/unlock it.
> Unfortunately, the error paths of alloc_uid() were not patched
> to unlock it.
> 

Thanks for catching it. Looks good to me. Ingo, could you please pick
these up?

> Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Acked-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>

-- 
regards,
Dhaval
Re: [PATCH 1/2] Don't forget to unlock uids_mutex on error paths [message #23730 is a reply to message #23728] Fri, 23 November 2007 13:52 Go to previous message
Ingo Molnar is currently offline  Ingo Molnar
Messages: 51
Registered: December 2005
Member
* Dhaval Giani <dhaval@linux.vnet.ibm.com> wrote:

> On Wed, Nov 21, 2007 at 01:49:33PM +0300, Pavel Emelyanov wrote:
> > The commit
> > 
> > commit 5cb350baf580017da38199625b7365b1763d7180
> > Author: Dhaval Giani <dhaval@linux.vnet.ibm.com>
> > Date:   Mon Oct 15 17:00:14 2007 +0200
> > 
> >     sched: group scheduling, sysfs tunables
> > 
> > introduced the uids_mutex and the helpers to lock/unlock it.
> > Unfortunately, the error paths of alloc_uid() were not patched
> > to unlock it.

oops!

> Thanks for catching it. Looks good to me. Ingo, could you please pick 
> these up?

yep - added.

	Ingo
Previous Topic: [PATCH][UNIX] Move the unix sock iterators in to proper place
Next Topic: [PATCH net-2.6.25] Name magic constants in sock_wake_async()
Goto Forum:
  


Current Time: Fri Dec 01 20:03:25 GMT 2023

Total time taken to generate the page: 0.04429 seconds