From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jon Maloy Subject: Re: skb_try_coalesce bug? Date: Tue, 22 Apr 2014 16:35:59 -0400 Message-ID: <5356D2AF.7030401@ericsson.com> References: <20140422120125.GC7019@eerihug-hybrid.rnd.ki.sw.ericsson.se> <1398172293.29946.37.camel@edumazet-glaptop2.roam.corp.google.com> <5356C518.7050602@ericsson.com> <1398197116.29946.60.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Erik Hugne , To: Eric Dumazet Return-path: Received: from usevmg20.ericsson.net ([198.24.6.45]:63519 "EHLO usevmg20.ericsson.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757594AbaDVUgL (ORCPT ); Tue, 22 Apr 2014 16:36:11 -0400 In-Reply-To: <1398197116.29946.60.camel@edumazet-glaptop2.roam.corp.google.com> Sender: netdev-owner@vger.kernel.org List-ID: On 04/22/2014 04:05 PM, Eric Dumazet wrote: > On Tue, 2014-04-22 at 15:38 -0400, Jon Maloy wrote: > >> >> In the case I encountered, our head buffer is linear (skb->data_len == 0), >> so it is the real tailroom value that is returned. An alas, that one is big >> enough to contain the last (small) fragment of the message. > > > Whole point of skb_try_coalesce() is to coalesce as much as possible, > without guarantee of keeping some sort of 'segments' > > skb_try_coalesce - try to merge skb to prior one > > If you do not want this to happen, (you seem to want nothing else in > your head buffer skb->head), you need to add some logic. Ok. I should have given a little background. 1: We send a message of 3041 bytes, inclusive TIPC header, via loopback interface. 2: This one gets chopped up in three fragments: 1420, 1420,and 201 bytes. (The mtu was of course wrong, but this is how I discovered the problem). 3: First fragment is received, uncloned, and serves as head. 4; Second fragment (a clone) is received. skb_try_coalesce() fails at the skb_head_is_locked() test, because the buffer is a clone. Because of this, we add the buffer to skb_shinfo(head)->frag_list instead. 5: Third fragment (also a clone) is received. Now, since we check for space in tailroom of header before we do anything else, it slips in there, and bypasses the already chained-up second segment. Regards ///jon > > A helper temporarily setting head->tail = head->end would do it I guess. > >