From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zoltan Kiss Subject: Re: [PATCH net RFC] xen-netback: Fix grant ref resolution in RX path Date: Wed, 14 May 2014 19:44:04 +0100 Message-ID: <5373B974.6050808@citrix.com> References: <1399991500-4432-1-git-send-email-zoltan.kiss@citrix.com> <1399997635.7973.56.camel@edumazet-glaptop2.roam.corp.google.com> <53734F9F.2010101@citrix.com> <1400067318.7973.69.camel@edumazet-glaptop2.roam.corp.google.com> <53736EBA.7050401@citrix.com> <1400085639.26934.17.camel@kazak.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: 7bit Cc: Eric Dumazet , , , , , , , To: Ian Campbell Return-path: Received: from smtp.citrix.com ([66.165.176.89]:2554 "EHLO SMTP.CITRIX.COM" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750750AbaENSo0 (ORCPT ); Wed, 14 May 2014 14:44:26 -0400 In-Reply-To: <1400085639.26934.17.camel@kazak.uk.xensource.com> Sender: netdev-owner@vger.kernel.org List-ID: On 14/05/14 17:40, Ian Campbell wrote: > > On Wed, 2014-05-14 at 14:25 +0100, Zoltan Kiss wrote: > >> But of course it should be carefully checked that functions which place >> frags into another frags arrays should call orphan_frags, e.g. I guess >> skb_shift does such thing. That's what I intend to start another thread >> about. > > I see Eric has posted a fix for skb_try_coalesce(), but is that likely > to be acceptable for 3.15? How many other similar patches are we > expecting and/or how long do you think this careful checking will take? > > Given that we are now at 3.15-rc5 I think we need to decide how to > proceed, either push ahead fixing all these issues or (partially) revert > this netback feature and try again for 3.16. I guess these problems apply to KVM as well, and it haven't caused them problem so far, so I guess the problem shouldn't be that big. I assume they are using this feature to send fragmented skbs out from a guest, but I might be wrong. Anyway, I think I'll check tomorrow if there are further places to worry about, but I don't think if there are any of them which are easily reproducible. Zoli