Re: [PATCH 1/2 v3] wait: add wait_event_lock_irq() interface

[Lukáš Czerner]

On Fri, 30 Nov 2012, Andrew Morton wrote:

> Date: Fri, 30 Nov 2012 12:13:08 -0800
> From: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> To: LukÃÅ Czerner <lczerner@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx, linux-raid@xxxxxxxxxxxxxxx, axboe@xxxxxxxxx,
>     jmoyer@xxxxxxxxxx, Neil Brown <neilb@xxxxxxx>
> Subject: Re: [PATCH 1/2 v3] wait: add wait_event_lock_irq() interface
> 
> On Fri, 30 Nov 2012 12:36:35 +0100 (CET)
> LukÃÅ Czerner <lczerner@xxxxxxxxxx> wrote:
> 
> > On Fri, 30 Nov 2012, Andrew Morton wrote:
> > 
> > > Date: Fri, 30 Nov 2012 02:48:42 -0800
> > > From: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> > > To: Lukas Czerner <lczerner@xxxxxxxxxx>
> > > Cc: linux-kernel@xxxxxxxxxxxxxxx, linux-raid@xxxxxxxxxxxxxxx, axboe@xxxxxxxxx,
> > >     jmoyer@xxxxxxxxxx, Neil Brown <neilb@xxxxxxx>,
> > >     David Howells <dhowells@xxxxxxxxxx>, Ingo Molnar <mingo@xxxxxxx>,
> > >     Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> > > Subject: Re: [PATCH 1/2 v3] wait: add wait_event_lock_irq() interface
> > > 
> > > On Fri, 30 Nov 2012 11:42:40 +0100 Lukas Czerner <lczerner@xxxxxxxxxx> wrote:
> > > 
> > > > New wait_event{_interruptible}_lock_irq{_cmd} macros added. This commit
> > > > moves the private wait_event_lock_irq() macro from MD to regular wait
> > > > includes, introduces new macro wait_event_lock_irq_cmd() instead of using
> > > > the old method with omitting cmd parameter which is ugly and makes a use
> > > > of new macros in the MD. It also introduces the _interruptible_ variant.
> > > > 
> > > > The use of new interface is when one have a special lock to protect data
> > > > structures used in the condition, or one also needs to invoke "cmd"
> > > > before putting it to sleep.
> > > > 
> > > > All new macros are expected to be called with the lock taken. The lock
> > > > is released before sleep and is reacquired afterwards. We will leave the
> > > > macro with the lock held.
> > > > 
> > > > Note to DM: IMO this should also fix theoretical race on waitqueue while
> > > > using simultaneously wait_event_lock_irq() and wait_event() because of
> > > > lack of locking around current state setting and wait queue removal.
> > > 
> > > Does this fix the sparse warning which Fengguang just sent us?
> > 
> > Which report from Fengguang do you have in mind ? I do not see any
> > on linux-kernel today.
> > 
> > /me going to see what spare reports
> > 
> 
> On Fri, 30 Nov 2012 16:30:24 +0800
> kbuild test robot <fengguang.wu@xxxxxxxxx> wrote:
> 
> > tree:   git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git akpm
> > head:   cfb65dadcd079ad4547407a1584bc6b96bd48bb3
> > commit: 2b29cdb6f98c86a1da4ec5335d6247392b7c6551 [35/476] wait: add wait_event_lock_irq() interface
> > 
> > 
> > sparse warnings:
> > 
> > + drivers/block/drbd/drbd_int.h:2339:9: sparse: preprocessor token __wait_event_lock_irq redefined
> > include/linux/wait.h:554:9: this was the original definition
> > + drivers/block/drbd/drbd_int.h:2358:9: sparse: preprocessor token wait_event_lock_irq redefined
> > include/linux/wait.h:621:9: this was the original definition
> > drivers/block/drbd/drbd_interval.h:12:63: sparse: dubious one-bit signed bitfield
> > drivers/block/drbd/drbd_interval.h:13:22: sparse: dubious one-bit signed bitfield
> > drivers/block/drbd/drbd_int.h:903:39: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1123:69: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1124:70: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1125:59: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1126:63: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1127:60: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1128:71: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1129:65: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1130:66: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1335:74: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1336:50: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1338:51: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1339:58: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1340:54: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1341:62: sparse: attribute 'require_context': unknown attribute
> > drivers/block/drbd/drbd_int.h:1435:92: sparse: attribute 'require_context': unknown attribute
> 

I believe that Lars send patch for that already. So no,

'[PATCH 1/2 v3] wait: add wait_event_lock_irq() interface'

does not fix the issue, but another patch should.

http://www.spinics.net/lists/linux-next/msg23042.html

Thanks!
-Lukas