From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Subject: Re: [PATCH] ext4: optimize ext4_end_io memory usage Date: Mon, 8 Apr 2013 23:38:44 +0200 Message-ID: <20130408213844.GC25878@quack.suse.cz> References: <1364793468-1302-1-git-send-email-dmonakhov@openvz.org> <20130405093514.GA3848@quack.suse.cz> <20130408171052.GB4598@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , Dmitry Monakhov , linux-ext4@vger.kernel.org To: Theodore Ts'o Return-path: Received: from cantor2.suse.de ([195.135.220.15]:53931 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S935728Ab3DHViq (ORCPT ); Mon, 8 Apr 2013 17:38:46 -0400 Content-Disposition: inline In-Reply-To: <20130408171052.GB4598@thunk.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Mon 08-04-13 13:10:52, Ted Tso wrote: > On Fri, Apr 05, 2013 at 11:35:14AM +0200, Jan Kara wrote: > > On Mon 01-04-13 09:17:48, Dmitry Monakhov wrote: > > > ext4_end_io->pages array has is used only for buffered writes and usless > > > in case of DIO. This patch allow us to save 1K for each DIO request. > > I have actually a more complete solution to this in my patch queue - > > there isn't any need to reference pages from io_end after my patches. > > Hopefully I'll be able to send things out later today... > > Hi Jan, > > Did you have a chance to send out your patches? Not on Friday but the patch bomb has landed just now (finally I was able to sort out all the failures in dioread_nolock mode). I'd just note that my version of Dmitry's patch is patch 1/29. Honza -- Jan Kara SUSE Labs, CR