Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-10 Thread Ilya Zykov
On 04.12.2012 11:07, Peter Hurley wrote: > > The most common cause stems from the n_tty_close() path spuriously > scheduling buffer work, when the ldisc has already been halted. > This is fixed in 'tty: Don't reschedule buffer work while closing' Thank you, very useful. Fix this: WARNING: at dr

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-06 Thread Peter Hurley
On Tue, 2012-12-04 at 02:07 -0500, Peter Hurley wrote: > This patch series addresses the causes of flush_to_ldisc accessing > the tty after freeing. Well, there's going to be a v2 of this series. Sasha found that PATCH 3/9 is an insufficient fix. I've already worked out the correct solution but I

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-04 Thread Alan Cox
> I assume you mean somewhere other than ml archives: > https://lkml.org/lkml/2012/12/1/57 Putting in the commit message would be idea given how short it is. > > Setting up a mini-testsuite is good idea. As I noted in that email > though, that test jig is a derivative work that I'd want a ok fro

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-04 Thread Peter Hurley
On Tue, 2012-12-04 at 08:54 +, Alan Cox wrote: > On Tue, 4 Dec 2012 02:07:36 -0500 > Peter Hurley wrote: > > > This patch series addresses the causes of flush_to_ldisc accessing > > the tty after freeing. > > Looks good to me. Would be nice to keep a copy of the test that shows > it up in t

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-04 Thread Jiri Slaby
On 12/04/2012 08:07 AM, Peter Hurley wrote: > This patch series addresses the causes of flush_to_ldisc accessing > the tty after freeing. Hi, thanks for doing the work. The series looks good to me. > This series also does not include Jiri's debug patch here > https://lkml.org/lkml/2012/11/2/278 f

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-04 Thread Alan Cox
On Tue, 4 Dec 2012 02:07:36 -0500 Peter Hurley wrote: > This patch series addresses the causes of flush_to_ldisc accessing > the tty after freeing. Looks good to me. Would be nice to keep a copy of the test that shows it up in the comments of the patches somewhere. -- To unsubscribe from this l

Re: [PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-03 Thread Ilya Zykov
On 04.12.2012 11:07, Peter Hurley wrote: > This patch series addresses the causes of flush_to_ldisc accessing > the tty after freeing. > I think, it is have sense only if you can take effect, with this patch or something like. I can't. :) Signed-off-by: Ilya Zykov --- diff --git a/drivers/tty/t

[PATCH -next 0/9] tty: Fix buffer work access-after-free

2012-12-03 Thread Peter Hurley
This patch series addresses the causes of flush_to_ldisc accessing the tty after freeing. The most common cause stems from the n_tty_close() path spuriously scheduling buffer work, when the ldisc has already been halted. This is fixed in 'tty: Don't reschedule buffer work while closing' The other