[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1261081167.27920.815.camel@laptop>
Date: Thu, 17 Dec 2009 21:19:27 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Christoph Lameter <cl@...ux-foundation.org>
Cc: Andi Kleen <andi@...stfloor.org>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"akpm@...ux-foundation.org" <akpm@...ux-foundation.org>,
"mingo@...e.hu" <mingo@...e.hu>, minchan.kim@...il.com
Subject: Re: [mm][RFC][PATCH 0/11] mm accessor updates.
On Thu, 2009-12-17 at 14:13 -0600, Christoph Lameter wrote:
> On Thu, 17 Dec 2009, Peter Zijlstra wrote:
>
> > > You always need some reference on the mm_struct (mm_read_lock) if you are
> > > going to sleep to ensure that mm_struct still exists after waking up (page
> > > fault, page allocation). RCU and other spin locks are not helping there.
> >
> > Depends what you go to sleep for, the page fault retry patches simply
> > retook the whole fault and there is no way the mm could have gone away
> > when userspace isn't executing.
>
> get_user_pages ?
on !current->mm that is? There's only very few cases for that, non of
which are on a hot-path iirc.
> > Also pinning a page will pin the vma will pin the mm, and then you can
> > always take explicit mm_struct refs, but you really want to avoid that
> > since that's a global cacheline again.
>
> Incrementing a refcount on some random page wont protect you unless
> mmap_sem is held.
Hmm, right, its pagetables I meant, but we can probably make it be true
for locking a page.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists