From mboxrd@z Thu Jan 1 00:00:00 1970 From: Artem Bityutskiy Subject: Re: [PATCHv6 00/15] kill unnecessary bdi wakeups + cleanups Date: Thu, 05 Aug 2010 12:35:47 +0300 Message-ID: <1281000947.15689.77.camel@localhost> References: <1280057365-10297-1-git-send-email-dedekind1@gmail.com> <4C580B1C.3020409@kernel.dk> <1280839036.15689.8.camel@localhost> <4C580FFF.2000605@kernel.dk> <4C595037.4050408@fusionio.com> Reply-To: dedekind1@gmail.com Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org To: Jens Axboe Return-path: Received: from smtp.nokia.com ([192.100.122.233]:49579 "EHLO mgw-mx06.nokia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759509Ab0HEJh3 (ORCPT ); Thu, 5 Aug 2010 05:37:29 -0400 In-Reply-To: <4C595037.4050408@fusionio.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Wed, 2010-08-04 at 13:34 +0200, Jens Axboe wrote: > On 2010-08-03 14:47, Jens Axboe wrote: > > On 2010-08-03 14:37, Artem Bityutskiy wrote: > >> On Tue, 2010-08-03 at 14:27 +0200, Jens Axboe wrote: > >>> On 2010-07-25 13:29, Artem Bityutskiy wrote: > >>>> Hi, > >>>> > >>>> here is v6 of the patch series which clean-ups bdi threads and s= ubstantially > >>>> lessens amount of unnecessary kernel wake-ups, which is very imp= ortant on > >>>> battery-powered devices. > >>>> > >>>> This patch-set is also available at: > >>>> git://git.infradead.org/users/dedekind/misc-2.6.git flushers_v6 > >>> > >>> Thanks Artem, for sticking around long enough to get this into > >>> shape. I have finally merged it. > >> > >> Thanks, but > >> > >>>> 1. Use 'spin_lock_bh' for the 'bdi->wb_lock' (changed patch N12) > >>> > >>> I'd rather not, question is how to avoid it. Either just wakeup t= he > >>> default thread, or punt the lock-and-check bdi->wb.task to a thre= ad. > >> > >> you merged this change, do you want me to send a separate patch wh= ich > >> undo the 'spin_lock_bh' change? I'll think about how to avoid this= and > >> come back. > >=20 > > Yes, it's not a huge thing, but it would be nice to get rid of. So = I > > figured it was better to merge it and not have you respin the serie= s yet > > again. >=20 > There is a spinlock bug in the current code, you nest _bh locks on lo= ck > but not always on unlock. I fixed it up as per the below: Right, sorry, to be frank I never used _bh spinlock versions before, an= d just did not know the nesting trick. Thanks for fixing. Also, I'm still not sure whether I should get rid of this _bh or not. I wrote my thoughts to you in another e-mail. --=20 Best Regards, Artem Bityutskiy (=D0=90=D1=80=D1=82=D1=91=D0=BC =D0=91=D0=B8=D1=82=D1=8E= =D1=86=D0=BA=D0=B8=D0=B9) -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel= " in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html