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 S1763161AbbA3WHu (ORCPT <rfc822;w@1wt.eu>);
	Fri, 30 Jan 2015 17:07:50 -0500
Received: from mail.linuxfoundation.org ([140.211.169.12]:47505 "EHLO
	mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753524AbbA3WHs (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Fri, 30 Jan 2015 17:07:48 -0500
Date: Fri, 30 Jan 2015 14:07:47 -0800
From: Greg KH <gregkh@linuxfoundation.org>
To: Tim Chen <tim.c.chen@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>, Akinobu Mita <akinobu.mita@gmail.com>,
        Mathias Nyman <mathias.nyman@linux.intel.com>,
        Andi Kleen <andi@firstfloor.org>, Ingo Molnar <mingo@elte.hu>,
        Andrew Morton <akpm@linux-foundation.org>,
        Marek Szyprowski <m.szyprowski@samsung.com>,
        Thomas Gleixner <tglx@linutronix.de>, linux-kernel@vger.kernel.org,
        x86@kernel.org, linux-usb@vger.kernel.org
Subject: Re: [PATCH] pci-dma: Fix x86 dma_alloc_coherent to fully clear all
 pages returned
Message-ID: <20150130220747.GA21792@kroah.com>
References: <1422647641.9530.2.camel@schen9-desk2.jf.intel.com>
 <20150130195853.GA17115@kroah.com>
 <1422655318.9530.15.camel@schen9-desk2.jf.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1422655318.9530.15.camel@schen9-desk2.jf.intel.com>
User-Agent: Mutt/1.5.23 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Fri, Jan 30, 2015 at 02:01:58PM -0800, Tim Chen wrote:
> On Fri, 2015-01-30 at 11:58 -0800, Greg KH wrote:
> 
> > 
> > Shouldn't this go to stable trees too?
> > 
> 
> Yes. Added the stable tag in updated patch that's resent.
> 
> > Also, why is the xhci driver not asking for the memory it is going to
> > need?  If it wants to use the full page, shouldn't it ask for it?
> > 
> 
> I agree that xhci should have done that, but it didn't.  Commit 
> d92ef66c4f8f ("x86: make dma_alloc_coherent() return zeroed memory 
> if CMA is enabled") changed the behavior of dma_alloc_coherent
> by clearing only the memory being asked for.
> 
> So for backward compatibility, clearing the pages 
> completely to revert to dma_alloc_coherent's original
> behavior is probably the safe thing to do.

We don't "need" any backward compatility, why not fix the broken drivers
that are using memory outside of what they are asking for?  That's not
ok no matter what, right?

thanks,

greg k-h