From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S934190Ab1KJJgR (ORCPT <rfc822;w@1wt.eu>);
	Thu, 10 Nov 2011 04:36:17 -0500
Received: from mx1.redhat.com ([209.132.183.28]:56024 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S934054Ab1KJJgP (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Thu, 10 Nov 2011 04:36:15 -0500
Date: Thu, 10 Nov 2011 10:34:42 +0100
From: Johannes Weiner <jweiner@redhat.com>
To: Andrea Arcangeli <aarcange@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>, Jan Kara <jack@suse.cz>,
        Andy Isaacson <adi@hexapodia.org>, linux-kernel@vger.kernel.org,
        linux-mm@vger.kernel.org
Subject: Re: long sleep_on_page delays writing to slow storage
Message-ID: <20111110093442.GG3153@redhat.com>
References: <20111107045928.GK8927@hexapodia.org>
 <20111109170027.GB7495@quack.suse.cz>
 <20111109175201.GB3083@suse.de>
 <20111109180646.GM5075@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20111109180646.GM5075@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Wed, Nov 09, 2011 at 07:06:46PM +0100, Andrea Arcangeli wrote:
> On Wed, Nov 09, 2011 at 05:52:01PM +0000, Mel Gorman wrote:
> > -#define alloc_pages_vma(gfp_mask, order, vma, addr, node)	\
> > -	alloc_pages(gfp_mask, order)
> > +#define alloc_pages_vma(gfp_mask, order, vma, addr, node, drop_mmapsem)	\
> > +	({ 								\
> > +		if (drop_mmapsem)					\
> > +			up_read(&vma->vm_mm->mmap_sem);			\
> > +		alloc_pages(gfp_mask, order);				\
> > +	})
> 
> I wouldn't change alloc_pages_vma. I think it's better to add and have
> that called only by khugepaged:
> 
> alloc_pages_vma_up_read(gfp_mask, order, vma, addr, node)
> {
> 	__alloc_pages_vma(gfp_mask, order, vma, addr, node, true);
> }
> 
> alloc_pages_vma(gfp_mask, order, vma, addr, node)
> {
> 	__alloc_pages_vma(gfp_mask, order, vma, addr, node, false);
> }
> 
> I wonder if a change like this would be enough?
> 
>        sync_migration = !(gfp_mask & __GFP_NO_KSWAPD);
> 
> But even if hidden in a new function, the main downside overall is the
> fact we'll pass one more var through the stack of fast paths.
> 
> Johannes I recall you reported this too and Mel suggested the above
> change, did it help in the end?

Yes, it completely fixed the latency problem.

That said, I haven't looked at the impact on the THP success rate, but
a regression there is probably less severe than half-minute-stalls in
interactive applications.