xfs-masters
[Top] [All Lists]

Re: RFC: Fix f_flags races without the BKL

To: Oleg Nesterov <oleg@xxxxxxxxxx>
Subject: Re: RFC: Fix f_flags races without the BKL
From: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Date: Fri, 2 Jan 2009 19:54:46 +0000
Cc: Jonathan Corbet <corbet@xxxxxxx>, LKML <linux-kernel@xxxxxxxxxxxxxxx>, Andi Kleen <andi@xxxxxxxxxxxxxx>, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>, bfields@xxxxxxxxxxxx, xfs-masters@xxxxxxxxxxx
In-reply-to: <20090102190902.GA25969@xxxxxxxxxx>
References: <20081229041352.6bbdf57c@tpl> <20090102184232.GH28946@xxxxxxxxxxxxxxxxxx> <20090102190902.GA25969@xxxxxxxxxx>
Sender: Al Viro <viro@xxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.17 (2007-11-01)
On Fri, Jan 02, 2009 at 08:09:03PM +0100, Oleg Nesterov wrote:
> On 01/02, Al Viro wrote:
> >
> > Bloating with mutex is over the top, indeed, but why can't we simply keep
> > a pointer to fasync_struct in there?  Do we ever have a struct file with
> > several fasync_struct?
> 
> pipe_rdwr_fasync() ?

Ho-hum...  Right you are ;-/

FWIW, it's still bloody tempting to try.  How about hlist from struct file
through fasync_struct?  Possibly with reference from fasync_struct back
to the queue it's on, while we are at it - would make fasync_helper simpler...

<Prev in Thread] Current Thread [Next in Thread>