Messages in this thread | | | From | David Howells <> | Subject | Re: [patch 7/9] signals: Fix more rcu assumptions | Date | Fri, 11 Dec 2009 13:59:04 +0000 |
| |
Thomas Gleixner <tglx@linutronix.de> wrote:
> > Perhaps it is better to modify __sigqueue_alloc() instead? It can take > > rcu_lock() around cred->user itself. > > Indeed. Was too tired to see the obvious :)
Ah, but... If __sigqueue_alloc() is called from sigqueue_alloc(), then you don't need the RCU read lock as the target task is current.
So perhaps the callsite for __sigqueue_alloc() in __send_signal()? That at least puts the rcu_read_lock() call in proximity to the function that actually needs it.
David
| |